Text Generation
Transformers
Safetensors
qwen4_exp
image-text-to-text
qwen
qwen4-exp
Mixture of Experts
nvfp4
modelopt
mtp
speculative-decoding
cybersecurity
security-research
red-team
blue-team
purple-team
llm-agent
fuzzing
vulnerability-research
tool-use
conversational
8-bit precision
Instructions to use Blackfrost-AI/CYBER-FROST-3.8-NVFP4-V2 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Transformers
How to use Blackfrost-AI/CYBER-FROST-3.8-NVFP4-V2 with Transformers:
# Use a pipeline as a high-level helper from transformers import pipeline pipe = pipeline("text-generation", model="Blackfrost-AI/CYBER-FROST-3.8-NVFP4-V2") messages = [ { "role": "user", "content": [ {"type": "image", "url": "https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/p-blog/candy.JPG"}, {"type": "text", "text": "What animal is on the candy?"} ] }, ] pipe(text=messages)# pip install -U transformers accelerate # Load model directly from transformers import AutoProcessor, AutoModelForMultimodalLM processor = AutoProcessor.from_pretrained("Blackfrost-AI/CYBER-FROST-3.8-NVFP4-V2") model = AutoModelForMultimodalLM.from_pretrained("Blackfrost-AI/CYBER-FROST-3.8-NVFP4-V2", device_map="auto") messages = [ { "role": "user", "content": [ {"type": "image", "url": "https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/p-blog/candy.JPG"}, {"type": "text", "text": "What animal is on the candy?"} ] }, ] inputs = processor.apply_chat_template( messages, add_generation_prompt=True, tokenize=True, return_dict=True, return_tensors="pt", ).to(model.device) outputs = model.generate(**inputs, max_new_tokens=256) print(processor.decode(outputs[0][inputs["input_ids"].shape[-1]:])) - Notebooks
- Google Colab
- Kaggle
- Local Apps Settings
- vLLM
How to use Blackfrost-AI/CYBER-FROST-3.8-NVFP4-V2 with vLLM:
Install from pip and serve model
# Install vLLM from pip: pip install vllm # Start the vLLM server: vllm serve "Blackfrost-AI/CYBER-FROST-3.8-NVFP4-V2" # Call the server using curl (OpenAI-compatible API): curl -X POST "http://localhost:8000/v1/chat/completions" \ -H "Content-Type: application/json" \ --data '{ "model": "Blackfrost-AI/CYBER-FROST-3.8-NVFP4-V2", "messages": [ { "role": "user", "content": "What is the capital of France?" } ] }'Use Docker
docker model run hf.co/Blackfrost-AI/CYBER-FROST-3.8-NVFP4-V2
- SGLang
How to use Blackfrost-AI/CYBER-FROST-3.8-NVFP4-V2 with SGLang:
Install from pip and serve model
# Install SGLang from pip: pip install sglang # Start the SGLang server: python3 -m sglang.launch_server \ --model-path "Blackfrost-AI/CYBER-FROST-3.8-NVFP4-V2" \ --host 0.0.0.0 \ --port 30000 # Call the server using curl (OpenAI-compatible API): curl -X POST "http://localhost:30000/v1/chat/completions" \ -H "Content-Type: application/json" \ --data '{ "model": "Blackfrost-AI/CYBER-FROST-3.8-NVFP4-V2", "messages": [ { "role": "user", "content": "What is the capital of France?" } ] }'Use Docker images
docker run --gpus all \ --shm-size 32g \ -p 30000:30000 \ -v ~/.cache/huggingface:/root/.cache/huggingface \ --env "HF_TOKEN=<secret>" \ --ipc=host \ lmsysorg/sglang:latest \ python3 -m sglang.launch_server \ --model-path "Blackfrost-AI/CYBER-FROST-3.8-NVFP4-V2" \ --host 0.0.0.0 \ --port 30000 # Call the server using curl (OpenAI-compatible API): curl -X POST "http://localhost:30000/v1/chat/completions" \ -H "Content-Type: application/json" \ --data '{ "model": "Blackfrost-AI/CYBER-FROST-3.8-NVFP4-V2", "messages": [ { "role": "user", "content": "What is the capital of France?" } ] }' - Docker Model Runner
How to use Blackfrost-AI/CYBER-FROST-3.8-NVFP4-V2 with Docker Model Runner:
docker model run hf.co/Blackfrost-AI/CYBER-FROST-3.8-NVFP4-V2
V2 model card: community feedback, gate/up scale mismatch and BF16 re-export correction
05b89d8 verified Download NVFP4_PROVENANCE.json from Blackfrost-AI/CYBER-FROST-3.8-NVFP4-V2: direct link, hf CLI and curl.
- Browser
- Download file 129 kB
-
https://huggingface.co/Blackfrost-AI/CYBER-FROST-3.8-NVFP4-V2/resolve/main/NVFP4_PROVENANCE.json
- Command line
-
hf download hf://Blackfrost-AI/CYBER-FROST-3.8-NVFP4-V2/NVFP4_PROVENANCE.json
-
curl -L -o NVFP4_PROVENANCE.json https://huggingface.co/Blackfrost-AI/CYBER-FROST-3.8-NVFP4-V2/resolve/main/NVFP4_PROVENANCE.json
129 kB
| { | |
| "calibration_inheritance": "input_scale tensors retained exactly from existing Cyber-Frost NVFP4 reference", | |
| "calibration_reference": "nvidia/Qwen3.8-Flash-Next-NVFP4", | |
| "calibration_reference_revision": "fc694b54fb0174e0913e6adf86691ef85a4ead47", | |
| "expert_batch_size": 32, | |
| "full_weight_hashes": "not performed", | |
| "gate_up_pairs_checked": 24576, | |
| "hardlink_note": "Hardlinked unchanged shards are regular files sharing source storage; never edit them in place.", | |
| "modelopt_commit": "022767c7ab3d7d36211affd85e5c496770cde768", | |
| "modelopt_version": "0.47.0rc0", | |
| "output": "/release/CYBER-FROST-3.8-NVFP4", | |
| "output_indexed_tensor_bytes": 186356367352, | |
| "output_keys": 296474, | |
| "output_shards": 131, | |
| "preserved_bf16": "all source non-routed tensors, including attention, shared experts, PLE, vision, MTP, embeddings, lm_head, norms, routers and hyperconnections", | |
| "quantization": "main routed experts only; ModelOpt NVFP4 W4A4 group16 max; per-expert gate/up common max BEFORE packing", | |
| "reference": "/reference", | |
| "schema": "blackfrost.cyber_frost_nvfp4.shared_w13.v1", | |
| "shards": [ | |
| { | |
| "calibrated_gate_up_inputs_equal": true, | |
| "calibrated_input_scales_preserved_exactly": true, | |
| "elapsed_seconds": 1.023, | |
| "emitted_expert_matrices": 0, | |
| "finite_positive_global_and_block_scales": true, | |
| "fused_source_tensors": 0, | |
| "gate_up_pairs_checked": 0, | |
| "output": "model-00001-of-00131.safetensors", | |
| "output_bytes": 1040155944, | |
| "output_keys": 349, | |
| "sample_reconstruction": {}, | |
| "schema": "blackfrost.cyber_frost_nvfp4.shared_w13.v1", | |
| "shard_number": 1, | |
| "source": "model-00001-of-00131.safetensors", | |
| "storage_mode": "copy_unchanged_cross_device" | |
| }, | |
| { | |
| "calibrated_gate_up_inputs_equal": true, | |
| "calibrated_input_scales_preserved_exactly": true, | |
| "elapsed_seconds": 14.258, | |
| "emitted_expert_matrices": 1024, | |
| "finite_positive_global_and_block_scales": true, | |
| "fused_source_tensors": 1, | |
| "gate_up_pairs_checked": 512, | |
| "output": "model-00002-of-00131.safetensors", | |
| "output_bytes": 944264728, | |
| "output_keys": 4096, | |
| "sample_reconstruction": { | |
| "gate_proj": { | |
| "absolute_error_sum": 5.7311073878081515, | |
| "error_squared_sum": 0.009431670759361167, | |
| "max_absolute_error": 0.0080217644572258, | |
| "sample_values": 6144, | |
| "source_squared_sum": 1.0850464425685553 | |
| }, | |
| "up_proj": { | |
| "absolute_error_sum": 5.600201637716964, | |
| "error_squared_sum": 0.009141445846925913, | |
| "max_absolute_error": 0.007498603314161301, | |
| "sample_values": 6144, | |
| "source_squared_sum": 1.0384334280451712 | |
| } | |
| }, | |
| "schema": "blackfrost.cyber_frost_nvfp4.shared_w13.v1", | |
| "shard_number": 2, | |
| "source": "model-00002-of-00131.safetensors", | |
| "storage_mode": "requantized_routed_experts_preserved_other_source_tensors" | |
| }, | |
| { | |
| "calibrated_gate_up_inputs_equal": true, | |
| "calibrated_input_scales_preserved_exactly": true, | |
| "elapsed_seconds": 4.19, | |
| "emitted_expert_matrices": 512, | |
| "finite_positive_global_and_block_scales": true, | |
| "fused_source_tensors": 1, | |
| "gate_up_pairs_checked": 0, | |
| "output": "model-00003-of-00131.safetensors", | |
| "output_bytes": 626931640, | |
| "output_keys": 2070, | |
| "sample_reconstruction": { | |
| "down_proj": { | |
| "absolute_error_sum": 5.6075335872592404, | |
| "error_squared_sum": 0.009395735511147332, | |
| "max_absolute_error": 0.0079636350274086, | |
| "sample_values": 6144, | |
| "source_squared_sum": 1.0233638619598076 | |
| } | |
| }, | |
| "schema": "blackfrost.cyber_frost_nvfp4.shared_w13.v1", | |
| "shard_number": 3, | |
| "source": "model-00003-of-00131.safetensors", | |
| "storage_mode": "requantized_routed_experts_preserved_other_source_tensors" | |
| }, | |
| { | |
| "calibrated_gate_up_inputs_equal": true, | |
| "calibrated_input_scales_preserved_exactly": true, | |
| "elapsed_seconds": 10.883, | |
| "emitted_expert_matrices": 1024, | |
| "finite_positive_global_and_block_scales": true, | |
| "fused_source_tensors": 1, | |
| "gate_up_pairs_checked": 512, | |
| "output": "model-00004-of-00131.safetensors", | |
| "output_bytes": 944264728, | |
| "output_keys": 4096, | |
| "sample_reconstruction": { | |
| "gate_proj": { | |
| "absolute_error_sum": 5.436450787237845, | |
| "error_squared_sum": 0.008520193664473225, | |
| "max_absolute_error": 0.005667552351951599, | |
| "sample_values": 6144, | |
| "source_squared_sum": 0.965168372529011 | |
| }, | |
| "up_proj": { | |
| "absolute_error_sum": 5.5543124484829605, | |
| "error_squared_sum": 0.009029990910940283, | |
| "max_absolute_error": 0.0071847084909677505, | |
| "sample_values": 6144, | |
| "source_squared_sum": 0.9964221583153194 | |
| } | |
| }, | |
| "schema": "blackfrost.cyber_frost_nvfp4.shared_w13.v1", | |
| "shard_number": 4, | |
| "source": "model-00004-of-00131.safetensors", | |
| "storage_mode": "requantized_routed_experts_preserved_other_source_tensors" | |
| }, | |
| { | |
| "calibrated_gate_up_inputs_equal": true, | |
| "calibrated_input_scales_preserved_exactly": true, | |
| "elapsed_seconds": 6.625, | |
| "emitted_expert_matrices": 512, | |
| "finite_positive_global_and_block_scales": true, | |
| "fused_source_tensors": 1, | |
| "gate_up_pairs_checked": 0, | |
| "output": "model-00005-of-00131.safetensors", | |
| "output_bytes": 2150303984, | |
| "output_keys": 2064, | |
| "sample_reconstruction": { | |
| "down_proj": { | |
| "absolute_error_sum": 5.326563806622289, | |
| "error_squared_sum": 0.008381249645750155, | |
| "max_absolute_error": 0.0064871646463871, | |
| "sample_values": 6144, | |
| "source_squared_sum": 0.8968875808871788 | |
| } | |
| }, | |
| "schema": "blackfrost.cyber_frost_nvfp4.shared_w13.v1", | |
| "shard_number": 5, | |
| "source": "model-00005-of-00131.safetensors", | |
| "storage_mode": "requantized_routed_experts_preserved_other_source_tensors" | |
| }, | |
| { | |
| "calibrated_gate_up_inputs_equal": true, | |
| "calibrated_input_scales_preserved_exactly": true, | |
| "elapsed_seconds": 6.622, | |
| "emitted_expert_matrices": 0, | |
| "finite_positive_global_and_block_scales": true, | |
| "fused_source_tensors": 0, | |
| "gate_up_pairs_checked": 0, | |
| "output": "model-00006-of-00131.safetensors", | |
| "output_bytes": 3200016032, | |
| "output_keys": 4, | |
| "sample_reconstruction": {}, | |
| "schema": "blackfrost.cyber_frost_nvfp4.shared_w13.v1", | |
| "shard_number": 6, | |
| "source": "model-00006-of-00131.safetensors", | |
| "storage_mode": "copy_unchanged_cross_device" | |
| }, | |
| { | |
| "calibrated_gate_up_inputs_equal": true, | |
| "calibrated_input_scales_preserved_exactly": true, | |
| "elapsed_seconds": 7.097, | |
| "emitted_expert_matrices": 0, | |
| "finite_positive_global_and_block_scales": true, | |
| "fused_source_tensors": 0, | |
| "gate_up_pairs_checked": 0, | |
| "output": "model-00007-of-00131.safetensors", | |
| "output_bytes": 3200016032, | |
| "output_keys": 4, | |
| "sample_reconstruction": {}, | |
| "schema": "blackfrost.cyber_frost_nvfp4.shared_w13.v1", | |
| "shard_number": 7, | |
| "source": "model-00007-of-00131.safetensors", | |
| "storage_mode": "copy_unchanged_cross_device" | |
| }, | |
| { | |
| "calibrated_gate_up_inputs_equal": true, | |
| "calibrated_input_scales_preserved_exactly": true, | |
| "elapsed_seconds": 7.703, | |
| "emitted_expert_matrices": 0, | |
| "finite_positive_global_and_block_scales": true, | |
| "fused_source_tensors": 0, | |
| "gate_up_pairs_checked": 0, | |
| "output": "model-00008-of-00131.safetensors", | |
| "output_bytes": 3200016032, | |
| "output_keys": 4, | |
| "sample_reconstruction": {}, | |
| "schema": "blackfrost.cyber_frost_nvfp4.shared_w13.v1", | |
| "shard_number": 8, | |
| "source": "model-00008-of-00131.safetensors", | |
| "storage_mode": "copy_unchanged_cross_device" | |
| }, | |
| { | |
| "calibrated_gate_up_inputs_equal": true, | |
| "calibrated_input_scales_preserved_exactly": true, | |
| "elapsed_seconds": 7.571, | |
| "emitted_expert_matrices": 0, | |
| "finite_positive_global_and_block_scales": true, | |
| "fused_source_tensors": 0, | |
| "gate_up_pairs_checked": 0, | |
| "output": "model-00009-of-00131.safetensors", | |
| "output_bytes": 3200016032, | |
| "output_keys": 4, | |
| "sample_reconstruction": {}, | |
| "schema": "blackfrost.cyber_frost_nvfp4.shared_w13.v1", | |
| "shard_number": 9, | |
| "source": "model-00009-of-00131.safetensors", | |
| "storage_mode": "copy_unchanged_cross_device" | |
| }, | |
| { | |
| "calibrated_gate_up_inputs_equal": true, | |
| "calibrated_input_scales_preserved_exactly": true, | |
| "elapsed_seconds": 7.611, | |
| "emitted_expert_matrices": 0, | |
| "finite_positive_global_and_block_scales": true, | |
| "fused_source_tensors": 0, | |
| "gate_up_pairs_checked": 0, | |
| "output": "model-00010-of-00131.safetensors", | |
| "output_bytes": 3200016032, | |
| "output_keys": 4, | |
| "sample_reconstruction": {}, | |
| "schema": "blackfrost.cyber_frost_nvfp4.shared_w13.v1", | |
| "shard_number": 10, | |
| "source": "model-00010-of-00131.safetensors", | |
| "storage_mode": "copy_unchanged_cross_device" | |
| }, | |
| { | |
| "calibrated_gate_up_inputs_equal": true, | |
| "calibrated_input_scales_preserved_exactly": true, | |
| "elapsed_seconds": 7.688, | |
| "emitted_expert_matrices": 0, | |
| "finite_positive_global_and_block_scales": true, | |
| "fused_source_tensors": 0, | |
| "gate_up_pairs_checked": 0, | |
| "output": "model-00011-of-00131.safetensors", | |
| "output_bytes": 3200016032, | |
| "output_keys": 4, | |
| "sample_reconstruction": {}, | |
| "schema": "blackfrost.cyber_frost_nvfp4.shared_w13.v1", | |
| "shard_number": 11, | |
| "source": "model-00011-of-00131.safetensors", | |
| "storage_mode": "copy_unchanged_cross_device" | |
| }, | |
| { | |
| "calibrated_gate_up_inputs_equal": true, | |
| "calibrated_input_scales_preserved_exactly": true, | |
| "elapsed_seconds": 7.637, | |
| "emitted_expert_matrices": 0, | |
| "finite_positive_global_and_block_scales": true, | |
| "fused_source_tensors": 0, | |
| "gate_up_pairs_checked": 0, | |
| "output": "model-00012-of-00131.safetensors", | |
| "output_bytes": 3200016032, | |
| "output_keys": 4, | |
| "sample_reconstruction": {}, | |
| "schema": "blackfrost.cyber_frost_nvfp4.shared_w13.v1", | |
| "shard_number": 12, | |
| "source": "model-00012-of-00131.safetensors", | |
| "storage_mode": "copy_unchanged_cross_device" | |
| }, | |
| { | |
| "calibrated_gate_up_inputs_equal": true, | |
| "calibrated_input_scales_preserved_exactly": true, | |
| "elapsed_seconds": 7.835, | |
| "emitted_expert_matrices": 0, | |
| "finite_positive_global_and_block_scales": true, | |
| "fused_source_tensors": 0, | |
| "gate_up_pairs_checked": 0, | |
| "output": "model-00013-of-00131.safetensors", | |
| "output_bytes": 3200016032, | |
| "output_keys": 4, | |
| "sample_reconstruction": {}, | |
| "schema": "blackfrost.cyber_frost_nvfp4.shared_w13.v1", | |
| "shard_number": 13, | |
| "source": "model-00013-of-00131.safetensors", | |
| "storage_mode": "copy_unchanged_cross_device" | |
| }, | |
| { | |
| "calibrated_gate_up_inputs_equal": true, | |
| "calibrated_input_scales_preserved_exactly": true, | |
| "elapsed_seconds": 7.676, | |
| "emitted_expert_matrices": 0, | |
| "finite_positive_global_and_block_scales": true, | |
| "fused_source_tensors": 0, | |
| "gate_up_pairs_checked": 0, | |
| "output": "model-00014-of-00131.safetensors", | |
| "output_bytes": 3200016032, | |
| "output_keys": 4, | |
| "sample_reconstruction": {}, | |
| "schema": "blackfrost.cyber_frost_nvfp4.shared_w13.v1", | |
| "shard_number": 14, | |
| "source": "model-00014-of-00131.safetensors", | |
| "storage_mode": "copy_unchanged_cross_device" | |
| }, | |
| { | |
| "calibrated_gate_up_inputs_equal": true, | |
| "calibrated_input_scales_preserved_exactly": true, | |
| "elapsed_seconds": 7.513, | |
| "emitted_expert_matrices": 0, | |
| "finite_positive_global_and_block_scales": true, | |
| "fused_source_tensors": 0, | |
| "gate_up_pairs_checked": 0, | |
| "output": "model-00015-of-00131.safetensors", | |
| "output_bytes": 3200016032, | |
| "output_keys": 4, | |
| "sample_reconstruction": {}, | |
| "schema": "blackfrost.cyber_frost_nvfp4.shared_w13.v1", | |
| "shard_number": 15, | |
| "source": "model-00015-of-00131.safetensors", | |
| "storage_mode": "copy_unchanged_cross_device" | |
| }, | |
| { | |
| "calibrated_gate_up_inputs_equal": true, | |
| "calibrated_input_scales_preserved_exactly": true, | |
| "elapsed_seconds": 7.377, | |
| "emitted_expert_matrices": 0, | |
| "finite_positive_global_and_block_scales": true, | |
| "fused_source_tensors": 0, | |
| "gate_up_pairs_checked": 0, | |
| "output": "model-00016-of-00131.safetensors", | |
| "output_bytes": 3200016032, | |
| "output_keys": 4, | |
| "sample_reconstruction": {}, | |
| "schema": "blackfrost.cyber_frost_nvfp4.shared_w13.v1", | |
| "shard_number": 16, | |
| "source": "model-00016-of-00131.safetensors", | |
| "storage_mode": "copy_unchanged_cross_device" | |
| }, | |
| { | |
| "calibrated_gate_up_inputs_equal": true, | |
| "calibrated_input_scales_preserved_exactly": true, | |
| "elapsed_seconds": 6.994, | |
| "emitted_expert_matrices": 0, | |
| "finite_positive_global_and_block_scales": true, | |
| "fused_source_tensors": 0, | |
| "gate_up_pairs_checked": 0, | |
| "output": "model-00017-of-00131.safetensors", | |
| "output_bytes": 3200016032, | |
| "output_keys": 4, | |
| "sample_reconstruction": {}, | |
| "schema": "blackfrost.cyber_frost_nvfp4.shared_w13.v1", | |
| "shard_number": 17, | |
| "source": "model-00017-of-00131.safetensors", | |
| "storage_mode": "copy_unchanged_cross_device" | |
| }, | |
| { | |
| "calibrated_gate_up_inputs_equal": true, | |
| "calibrated_input_scales_preserved_exactly": true, | |
| "elapsed_seconds": 7.334, | |
| "emitted_expert_matrices": 0, | |
| "finite_positive_global_and_block_scales": true, | |
| "fused_source_tensors": 0, | |
| "gate_up_pairs_checked": 0, | |
| "output": "model-00018-of-00131.safetensors", | |
| "output_bytes": 3200016032, | |
| "output_keys": 4, | |
| "sample_reconstruction": {}, | |
| "schema": "blackfrost.cyber_frost_nvfp4.shared_w13.v1", | |
| "shard_number": 18, | |
| "source": "model-00018-of-00131.safetensors", | |
| "storage_mode": "copy_unchanged_cross_device" | |
| }, | |
| { | |
| "calibrated_gate_up_inputs_equal": true, | |
| "calibrated_input_scales_preserved_exactly": true, | |
| "elapsed_seconds": 7.522, | |
| "emitted_expert_matrices": 0, | |
| "finite_positive_global_and_block_scales": true, | |
| "fused_source_tensors": 0, | |
| "gate_up_pairs_checked": 0, | |
| "output": "model-00019-of-00131.safetensors", | |
| "output_bytes": 3200016032, | |
| "output_keys": 4, | |
| "sample_reconstruction": {}, | |
| "schema": "blackfrost.cyber_frost_nvfp4.shared_w13.v1", | |
| "shard_number": 19, | |
| "source": "model-00019-of-00131.safetensors", | |
| "storage_mode": "copy_unchanged_cross_device" | |
| }, | |
| { | |
| "calibrated_gate_up_inputs_equal": true, | |
| "calibrated_input_scales_preserved_exactly": true, | |
| "elapsed_seconds": 7.437, | |
| "emitted_expert_matrices": 0, | |
| "finite_positive_global_and_block_scales": true, | |
| "fused_source_tensors": 0, | |
| "gate_up_pairs_checked": 0, | |
| "output": "model-00020-of-00131.safetensors", | |
| "output_bytes": 3200016032, | |
| "output_keys": 4, | |
| "sample_reconstruction": {}, | |
| "schema": "blackfrost.cyber_frost_nvfp4.shared_w13.v1", | |
| "shard_number": 20, | |
| "source": "model-00020-of-00131.safetensors", | |
| "storage_mode": "copy_unchanged_cross_device" | |
| }, | |
| { | |
| "calibrated_gate_up_inputs_equal": true, | |
| "calibrated_input_scales_preserved_exactly": true, | |
| "elapsed_seconds": 7.684, | |
| "emitted_expert_matrices": 0, | |
| "finite_positive_global_and_block_scales": true, | |
| "fused_source_tensors": 0, | |
| "gate_up_pairs_checked": 0, | |
| "output": "model-00021-of-00131.safetensors", | |
| "output_bytes": 3200016032, | |
| "output_keys": 4, | |
| "sample_reconstruction": {}, | |
| "schema": "blackfrost.cyber_frost_nvfp4.shared_w13.v1", | |
| "shard_number": 21, | |
| "source": "model-00021-of-00131.safetensors", | |
| "storage_mode": "copy_unchanged_cross_device" | |
| }, | |
| { | |
| "calibrated_gate_up_inputs_equal": true, | |
| "calibrated_input_scales_preserved_exactly": true, | |
| "elapsed_seconds": 7.622, | |
| "emitted_expert_matrices": 0, | |
| "finite_positive_global_and_block_scales": true, | |
| "fused_source_tensors": 0, | |
| "gate_up_pairs_checked": 0, | |
| "output": "model-00022-of-00131.safetensors", | |
| "output_bytes": 3200016032, | |
| "output_keys": 4, | |
| "sample_reconstruction": {}, | |
| "schema": "blackfrost.cyber_frost_nvfp4.shared_w13.v1", | |
| "shard_number": 22, | |
| "source": "model-00022-of-00131.safetensors", | |
| "storage_mode": "copy_unchanged_cross_device" | |
| }, | |
| { | |
| "calibrated_gate_up_inputs_equal": true, | |
| "calibrated_input_scales_preserved_exactly": true, | |
| "elapsed_seconds": 7.513, | |
| "emitted_expert_matrices": 0, | |
| "finite_positive_global_and_block_scales": true, | |
| "fused_source_tensors": 0, | |
| "gate_up_pairs_checked": 0, | |
| "output": "model-00023-of-00131.safetensors", | |
| "output_bytes": 3200016032, | |
| "output_keys": 4, | |
| "sample_reconstruction": {}, | |
| "schema": "blackfrost.cyber_frost_nvfp4.shared_w13.v1", | |
| "shard_number": 23, | |
| "source": "model-00023-of-00131.safetensors", | |
| "storage_mode": "copy_unchanged_cross_device" | |
| }, | |
| { | |
| "calibrated_gate_up_inputs_equal": true, | |
| "calibrated_input_scales_preserved_exactly": true, | |
| "elapsed_seconds": 7.746, | |
| "emitted_expert_matrices": 0, | |
| "finite_positive_global_and_block_scales": true, | |
| "fused_source_tensors": 0, | |
| "gate_up_pairs_checked": 0, | |
| "output": "model-00024-of-00131.safetensors", | |
| "output_bytes": 3200016032, | |
| "output_keys": 4, | |
| "sample_reconstruction": {}, | |
| "schema": "blackfrost.cyber_frost_nvfp4.shared_w13.v1", | |
| "shard_number": 24, | |
| "source": "model-00024-of-00131.safetensors", | |
| "storage_mode": "copy_unchanged_cross_device" | |
| }, | |
| { | |
| "calibrated_gate_up_inputs_equal": true, | |
| "calibrated_input_scales_preserved_exactly": true, | |
| "elapsed_seconds": 7.656, | |
| "emitted_expert_matrices": 0, | |
| "finite_positive_global_and_block_scales": true, | |
| "fused_source_tensors": 0, | |
| "gate_up_pairs_checked": 0, | |
| "output": "model-00025-of-00131.safetensors", | |
| "output_bytes": 3200016032, | |
| "output_keys": 4, | |
| "sample_reconstruction": {}, | |
| "schema": "blackfrost.cyber_frost_nvfp4.shared_w13.v1", | |
| "shard_number": 25, | |
| "source": "model-00025-of-00131.safetensors", | |
| "storage_mode": "copy_unchanged_cross_device" | |
| }, | |
| { | |
| "calibrated_gate_up_inputs_equal": true, | |
| "calibrated_input_scales_preserved_exactly": true, | |
| "elapsed_seconds": 7.693, | |
| "emitted_expert_matrices": 0, | |
| "finite_positive_global_and_block_scales": true, | |
| "fused_source_tensors": 0, | |
| "gate_up_pairs_checked": 0, | |
| "output": "model-00026-of-00131.safetensors", | |
| "output_bytes": 3200016032, | |
| "output_keys": 4, | |
| "sample_reconstruction": {}, | |
| "schema": "blackfrost.cyber_frost_nvfp4.shared_w13.v1", | |
| "shard_number": 26, | |
| "source": "model-00026-of-00131.safetensors", | |
| "storage_mode": "copy_unchanged_cross_device" | |
| }, | |
| { | |
| "calibrated_gate_up_inputs_equal": true, | |
| "calibrated_input_scales_preserved_exactly": true, | |
| "elapsed_seconds": 7.625, | |
| "emitted_expert_matrices": 0, | |
| "finite_positive_global_and_block_scales": true, | |
| "fused_source_tensors": 0, | |
| "gate_up_pairs_checked": 0, | |
| "output": "model-00027-of-00131.safetensors", | |
| "output_bytes": 3200016032, | |
| "output_keys": 4, | |
| "sample_reconstruction": {}, | |
| "schema": "blackfrost.cyber_frost_nvfp4.shared_w13.v1", | |
| "shard_number": 27, | |
| "source": "model-00027-of-00131.safetensors", | |
| "storage_mode": "copy_unchanged_cross_device" | |
| }, | |
| { | |
| "calibrated_gate_up_inputs_equal": true, | |
| "calibrated_input_scales_preserved_exactly": true, | |
| "elapsed_seconds": 7.679, | |
| "emitted_expert_matrices": 0, | |
| "finite_positive_global_and_block_scales": true, | |
| "fused_source_tensors": 0, | |
| "gate_up_pairs_checked": 0, | |
| "output": "model-00028-of-00131.safetensors", | |
| "output_bytes": 3200016032, | |
| "output_keys": 4, | |
| "sample_reconstruction": {}, | |
| "schema": "blackfrost.cyber_frost_nvfp4.shared_w13.v1", | |
| "shard_number": 28, | |
| "source": "model-00028-of-00131.safetensors", | |
| "storage_mode": "copy_unchanged_cross_device" | |
| }, | |
| { | |
| "calibrated_gate_up_inputs_equal": true, | |
| "calibrated_input_scales_preserved_exactly": true, | |
| "elapsed_seconds": 7.568, | |
| "emitted_expert_matrices": 0, | |
| "finite_positive_global_and_block_scales": true, | |
| "fused_source_tensors": 0, | |
| "gate_up_pairs_checked": 0, | |
| "output": "model-00029-of-00131.safetensors", | |
| "output_bytes": 3200016032, | |
| "output_keys": 4, | |
| "sample_reconstruction": {}, | |
| "schema": "blackfrost.cyber_frost_nvfp4.shared_w13.v1", | |
| "shard_number": 29, | |
| "source": "model-00029-of-00131.safetensors", | |
| "storage_mode": "copy_unchanged_cross_device" | |
| }, | |
| { | |
| "calibrated_gate_up_inputs_equal": true, | |
| "calibrated_input_scales_preserved_exactly": true, | |
| "elapsed_seconds": 7.651, | |
| "emitted_expert_matrices": 0, | |
| "finite_positive_global_and_block_scales": true, | |
| "fused_source_tensors": 0, | |
| "gate_up_pairs_checked": 0, | |
| "output": "model-00030-of-00131.safetensors", | |
| "output_bytes": 3200016032, | |
| "output_keys": 4, | |
| "sample_reconstruction": {}, | |
| "schema": "blackfrost.cyber_frost_nvfp4.shared_w13.v1", | |
| "shard_number": 30, | |
| "source": "model-00030-of-00131.safetensors", | |
| "storage_mode": "copy_unchanged_cross_device" | |
| }, | |
| { | |
| "calibrated_gate_up_inputs_equal": true, | |
| "calibrated_input_scales_preserved_exactly": true, | |
| "elapsed_seconds": 7.683, | |
| "emitted_expert_matrices": 0, | |
| "finite_positive_global_and_block_scales": true, | |
| "fused_source_tensors": 0, | |
| "gate_up_pairs_checked": 0, | |
| "output": "model-00031-of-00131.safetensors", | |
| "output_bytes": 3200016040, | |
| "output_keys": 4, | |
| "sample_reconstruction": {}, | |
| "schema": "blackfrost.cyber_frost_nvfp4.shared_w13.v1", | |
| "shard_number": 31, | |
| "source": "model-00031-of-00131.safetensors", | |
| "storage_mode": "copy_unchanged_cross_device" | |
| }, | |
| { | |
| "calibrated_gate_up_inputs_equal": true, | |
| "calibrated_input_scales_preserved_exactly": true, | |
| "elapsed_seconds": 7.812, | |
| "emitted_expert_matrices": 0, | |
| "finite_positive_global_and_block_scales": true, | |
| "fused_source_tensors": 0, | |
| "gate_up_pairs_checked": 0, | |
| "output": "model-00032-of-00131.safetensors", | |
| "output_bytes": 3200016040, | |
| "output_keys": 4, | |
| "sample_reconstruction": {}, | |
| "schema": "blackfrost.cyber_frost_nvfp4.shared_w13.v1", | |
| "shard_number": 32, | |
| "source": "model-00032-of-00131.safetensors", | |
| "storage_mode": "copy_unchanged_cross_device" | |
| }, | |
| { | |
| "calibrated_gate_up_inputs_equal": true, | |
| "calibrated_input_scales_preserved_exactly": true, | |
| "elapsed_seconds": 7.778, | |
| "emitted_expert_matrices": 0, | |
| "finite_positive_global_and_block_scales": true, | |
| "fused_source_tensors": 0, | |
| "gate_up_pairs_checked": 0, | |
| "output": "model-00033-of-00131.safetensors", | |
| "output_bytes": 3200016040, | |
| "output_keys": 4, | |
| "sample_reconstruction": {}, | |
| "schema": "blackfrost.cyber_frost_nvfp4.shared_w13.v1", | |
| "shard_number": 33, | |
| "source": "model-00033-of-00131.safetensors", | |
| "storage_mode": "copy_unchanged_cross_device" | |
| }, | |
| { | |
| "calibrated_gate_up_inputs_equal": true, | |
| "calibrated_input_scales_preserved_exactly": true, | |
| "elapsed_seconds": 7.74, | |
| "emitted_expert_matrices": 0, | |
| "finite_positive_global_and_block_scales": true, | |
| "fused_source_tensors": 0, | |
| "gate_up_pairs_checked": 0, | |
| "output": "model-00034-of-00131.safetensors", | |
| "output_bytes": 3200016040, | |
| "output_keys": 4, | |
| "sample_reconstruction": {}, | |
| "schema": "blackfrost.cyber_frost_nvfp4.shared_w13.v1", | |
| "shard_number": 34, | |
| "source": "model-00034-of-00131.safetensors", | |
| "storage_mode": "copy_unchanged_cross_device" | |
| }, | |
| { | |
| "calibrated_gate_up_inputs_equal": true, | |
| "calibrated_input_scales_preserved_exactly": true, | |
| "elapsed_seconds": 8.272, | |
| "emitted_expert_matrices": 0, | |
| "finite_positive_global_and_block_scales": true, | |
| "fused_source_tensors": 0, | |
| "gate_up_pairs_checked": 0, | |
| "output": "model-00035-of-00131.safetensors", | |
| "output_bytes": 3200016040, | |
| "output_keys": 4, | |
| "sample_reconstruction": {}, | |
| "schema": "blackfrost.cyber_frost_nvfp4.shared_w13.v1", | |
| "shard_number": 35, | |
| "source": "model-00035-of-00131.safetensors", | |
| "storage_mode": "copy_unchanged_cross_device" | |
| }, | |
| { | |
| "calibrated_gate_up_inputs_equal": true, | |
| "calibrated_input_scales_preserved_exactly": true, | |
| "elapsed_seconds": 8.209, | |
| "emitted_expert_matrices": 0, | |
| "finite_positive_global_and_block_scales": true, | |
| "fused_source_tensors": 0, | |
| "gate_up_pairs_checked": 0, | |
| "output": "model-00036-of-00131.safetensors", | |
| "output_bytes": 3200016040, | |
| "output_keys": 4, | |
| "sample_reconstruction": {}, | |
| "schema": "blackfrost.cyber_frost_nvfp4.shared_w13.v1", | |
| "shard_number": 36, | |
| "source": "model-00036-of-00131.safetensors", | |
| "storage_mode": "copy_unchanged_cross_device" | |
| }, | |
| { | |
| "calibrated_gate_up_inputs_equal": true, | |
| "calibrated_input_scales_preserved_exactly": true, | |
| "elapsed_seconds": 4.599, | |
| "emitted_expert_matrices": 0, | |
| "finite_positive_global_and_block_scales": true, | |
| "fused_source_tensors": 0, | |
| "gate_up_pairs_checked": 0, | |
| "output": "model-00037-of-00131.safetensors", | |
| "output_bytes": 1742347160, | |
| "output_keys": 20, | |
| "sample_reconstruction": {}, | |
| "schema": "blackfrost.cyber_frost_nvfp4.shared_w13.v1", | |
| "shard_number": 37, | |
| "source": "model-00037-of-00131.safetensors", | |
| "storage_mode": "copy_unchanged_cross_device" | |
| }, | |
| { | |
| "calibrated_gate_up_inputs_equal": true, | |
| "calibrated_input_scales_preserved_exactly": true, | |
| "elapsed_seconds": 11.4, | |
| "emitted_expert_matrices": 1024, | |
| "finite_positive_global_and_block_scales": true, | |
| "fused_source_tensors": 1, | |
| "gate_up_pairs_checked": 512, | |
| "output": "model-00038-of-00131.safetensors", | |
| "output_bytes": 944268824, | |
| "output_keys": 4096, | |
| "sample_reconstruction": { | |
| "gate_proj": { | |
| "absolute_error_sum": 5.636302922386676, | |
| "error_squared_sum": 0.009018220497497032, | |
| "max_absolute_error": 0.0061907097697257996, | |
| "sample_values": 6144, | |
| "source_squared_sum": 1.031074805125083 | |
| }, | |
| "up_proj": { | |
| "absolute_error_sum": 5.645298554096371, | |
| "error_squared_sum": 0.009218631579164962, | |
| "max_absolute_error": 0.0075334832072257996, | |
| "sample_values": 6144, | |
| "source_squared_sum": 1.0009530297747773 | |
| } | |
| }, | |
| "schema": "blackfrost.cyber_frost_nvfp4.shared_w13.v1", | |
| "shard_number": 38, | |
| "source": "model-00038-of-00131.safetensors", | |
| "storage_mode": "requantized_routed_experts_preserved_other_source_tensors" | |
| }, | |
| { | |
| "calibrated_gate_up_inputs_equal": true, | |
| "calibrated_input_scales_preserved_exactly": true, | |
| "elapsed_seconds": 4.236, | |
| "emitted_expert_matrices": 512, | |
| "finite_positive_global_and_block_scales": true, | |
| "fused_source_tensors": 1, | |
| "gate_up_pairs_checked": 0, | |
| "output": "model-00039-of-00131.safetensors", | |
| "output_bytes": 511014256, | |
| "output_keys": 2061, | |
| "sample_reconstruction": { | |
| "down_proj": { | |
| "absolute_error_sum": 5.482149287126958, | |
| "error_squared_sum": 0.0086961299097449, | |
| "max_absolute_error": 0.005545482039451599, | |
| "sample_values": 6144, | |
| "source_squared_sum": 0.9645405514582548 | |
| } | |
| }, | |
| "schema": "blackfrost.cyber_frost_nvfp4.shared_w13.v1", | |
| "shard_number": 39, | |
| "source": "model-00039-of-00131.safetensors", | |
| "storage_mode": "requantized_routed_experts_preserved_other_source_tensors" | |
| }, | |
| { | |
| "calibrated_gate_up_inputs_equal": true, | |
| "calibrated_input_scales_preserved_exactly": true, | |
| "elapsed_seconds": 10.474, | |
| "emitted_expert_matrices": 1024, | |
| "finite_positive_global_and_block_scales": true, | |
| "fused_source_tensors": 1, | |
| "gate_up_pairs_checked": 512, | |
| "output": "model-00040-of-00131.safetensors", | |
| "output_bytes": 944268824, | |
| "output_keys": 4096, | |
| "sample_reconstruction": { | |
| "gate_proj": { | |
| "absolute_error_sum": 5.420939550967887, | |
| "error_squared_sum": 0.008449084669313676, | |
| "max_absolute_error": 0.0058942511677742004, | |
| "sample_values": 6144, | |
| "source_squared_sum": 0.9163162506382145 | |
| }, | |
| "up_proj": { | |
| "absolute_error_sum": 5.467191289237235, | |
| "error_squared_sum": 0.008624160231728115, | |
| "max_absolute_error": 0.0057547427713871, | |
| "sample_values": 6144, | |
| "source_squared_sum": 0.9554865126895926 | |
| } | |
| }, | |
| "schema": "blackfrost.cyber_frost_nvfp4.shared_w13.v1", | |
| "shard_number": 40, | |
| "source": "model-00040-of-00131.safetensors", | |
| "storage_mode": "requantized_routed_experts_preserved_other_source_tensors" | |
| }, | |
| { | |
| "calibrated_gate_up_inputs_equal": true, | |
| "calibrated_input_scales_preserved_exactly": true, | |
| "elapsed_seconds": 4.234, | |
| "emitted_expert_matrices": 512, | |
| "finite_positive_global_and_block_scales": true, | |
| "fused_source_tensors": 1, | |
| "gate_up_pairs_checked": 0, | |
| "output": "model-00041-of-00131.safetensors", | |
| "output_bytes": 736381704, | |
| "output_keys": 2080, | |
| "sample_reconstruction": { | |
| "down_proj": { | |
| "absolute_error_sum": 5.579542188788764, | |
| "error_squared_sum": 0.009001386516685316, | |
| "max_absolute_error": 0.00634765625, | |
| "sample_values": 6144, | |
| "source_squared_sum": 0.9773769645111172 | |
| } | |
| }, | |
| "schema": "blackfrost.cyber_frost_nvfp4.shared_w13.v1", | |
| "shard_number": 41, | |
| "source": "model-00041-of-00131.safetensors", | |
| "storage_mode": "requantized_routed_experts_preserved_other_source_tensors" | |
| }, | |
| { | |
| "calibrated_gate_up_inputs_equal": true, | |
| "calibrated_input_scales_preserved_exactly": true, | |
| "elapsed_seconds": 10.377, | |
| "emitted_expert_matrices": 1024, | |
| "finite_positive_global_and_block_scales": true, | |
| "fused_source_tensors": 1, | |
| "gate_up_pairs_checked": 512, | |
| "output": "model-00042-of-00131.safetensors", | |
| "output_bytes": 944268824, | |
| "output_keys": 4096, | |
| "sample_reconstruction": { | |
| "gate_proj": { | |
| "absolute_error_sum": 5.66801721137017, | |
| "error_squared_sum": 0.009160037436379682, | |
| "max_absolute_error": 0.005900062620639801, | |
| "sample_values": 6144, | |
| "source_squared_sum": 1.0124848505313149 | |
| }, | |
| "up_proj": { | |
| "absolute_error_sum": 5.601023465045728, | |
| "error_squared_sum": 0.009054554829875638, | |
| "max_absolute_error": 0.0059814453125, | |
| "sample_values": 6144, | |
| "source_squared_sum": 1.0110547533696814 | |
| } | |
| }, | |
| "schema": "blackfrost.cyber_frost_nvfp4.shared_w13.v1", | |
| "shard_number": 42, | |
| "source": "model-00042-of-00131.safetensors", | |
| "storage_mode": "requantized_routed_experts_preserved_other_source_tensors" | |
| }, | |
| { | |
| "calibrated_gate_up_inputs_equal": true, | |
| "calibrated_input_scales_preserved_exactly": true, | |
| "elapsed_seconds": 4.381, | |
| "emitted_expert_matrices": 512, | |
| "finite_positive_global_and_block_scales": true, | |
| "fused_source_tensors": 1, | |
| "gate_up_pairs_checked": 0, | |
| "output": "model-00043-of-00131.safetensors", | |
| "output_bytes": 633487440, | |
| "output_keys": 2071, | |
| "sample_reconstruction": { | |
| "down_proj": { | |
| "absolute_error_sum": 5.71136607893277, | |
| "error_squared_sum": 0.009335086122242934, | |
| "max_absolute_error": 0.0060163214802742004, | |
| "sample_values": 6144, | |
| "source_squared_sum": 1.0253474584864943 | |
| } | |
| }, | |
| "schema": "blackfrost.cyber_frost_nvfp4.shared_w13.v1", | |
| "shard_number": 43, | |
| "source": "model-00043-of-00131.safetensors", | |
| "storage_mode": "requantized_routed_experts_preserved_other_source_tensors" | |
| }, | |
| { | |
| "calibrated_gate_up_inputs_equal": true, | |
| "calibrated_input_scales_preserved_exactly": true, | |
| "elapsed_seconds": 11.346, | |
| "emitted_expert_matrices": 1024, | |
| "finite_positive_global_and_block_scales": true, | |
| "fused_source_tensors": 1, | |
| "gate_up_pairs_checked": 512, | |
| "output": "model-00044-of-00131.safetensors", | |
| "output_bytes": 944268824, | |
| "output_keys": 4096, | |
| "sample_reconstruction": { | |
| "gate_proj": { | |
| "absolute_error_sum": 5.35940574365668, | |
| "error_squared_sum": 0.008272196737227481, | |
| "max_absolute_error": 0.0058942511677742004, | |
| "sample_values": 6144, | |
| "source_squared_sum": 0.9602945761927251 | |
| }, | |
| "up_proj": { | |
| "absolute_error_sum": 5.522928553982638, | |
| "error_squared_sum": 0.008655439391432655, | |
| "max_absolute_error": 0.0058186836540699005, | |
| "sample_values": 6144, | |
| "source_squared_sum": 0.9852734779722375 | |
| } | |
| }, | |
| "schema": "blackfrost.cyber_frost_nvfp4.shared_w13.v1", | |
| "shard_number": 44, | |
| "source": "model-00044-of-00131.safetensors", | |
| "storage_mode": "requantized_routed_experts_preserved_other_source_tensors" | |
| }, | |
| { | |
| "calibrated_gate_up_inputs_equal": true, | |
| "calibrated_input_scales_preserved_exactly": true, | |
| "elapsed_seconds": 4.341, | |
| "emitted_expert_matrices": 512, | |
| "finite_positive_global_and_block_scales": true, | |
| "fused_source_tensors": 1, | |
| "gate_up_pairs_checked": 0, | |
| "output": "model-00045-of-00131.safetensors", | |
| "output_bytes": 626933712, | |
| "output_keys": 2070, | |
| "sample_reconstruction": { | |
| "down_proj": { | |
| "absolute_error_sum": 5.42722690559458, | |
| "error_squared_sum": 0.008562844664239356, | |
| "max_absolute_error": 0.0070219486951828, | |
| "sample_values": 6144, | |
| "source_squared_sum": 0.9509799469235858 | |
| } | |
| }, | |
| "schema": "blackfrost.cyber_frost_nvfp4.shared_w13.v1", | |
| "shard_number": 45, | |
| "source": "model-00045-of-00131.safetensors", | |
| "storage_mode": "requantized_routed_experts_preserved_other_source_tensors" | |
| }, | |
| { | |
| "calibrated_gate_up_inputs_equal": true, | |
| "calibrated_input_scales_preserved_exactly": true, | |
| "elapsed_seconds": 10.521, | |
| "emitted_expert_matrices": 1024, | |
| "finite_positive_global_and_block_scales": true, | |
| "fused_source_tensors": 1, | |
| "gate_up_pairs_checked": 512, | |
| "output": "model-00046-of-00131.safetensors", | |
| "output_bytes": 944268824, | |
| "output_keys": 4096, | |
| "sample_reconstruction": { | |
| "gate_proj": { | |
| "absolute_error_sum": 5.476230456784833, | |
| "error_squared_sum": 0.008669457568349366, | |
| "max_absolute_error": 0.006295341998338699, | |
| "sample_values": 6144, | |
| "source_squared_sum": 0.9308192135610556 | |
| }, | |
| "up_proj": { | |
| "absolute_error_sum": 5.5650695820804685, | |
| "error_squared_sum": 0.008946927964075119, | |
| "max_absolute_error": 0.00676618330180645, | |
| "sample_values": 6144, | |
| "source_squared_sum": 0.958331221451296 | |
| } | |
| }, | |
| "schema": "blackfrost.cyber_frost_nvfp4.shared_w13.v1", | |
| "shard_number": 46, | |
| "source": "model-00046-of-00131.safetensors", | |
| "storage_mode": "requantized_routed_experts_preserved_other_source_tensors" | |
| }, | |
| { | |
| "calibrated_gate_up_inputs_equal": true, | |
| "calibrated_input_scales_preserved_exactly": true, | |
| "elapsed_seconds": 4.01, | |
| "emitted_expert_matrices": 512, | |
| "finite_positive_global_and_block_scales": true, | |
| "fused_source_tensors": 1, | |
| "gate_up_pairs_checked": 0, | |
| "output": "model-00047-of-00131.safetensors", | |
| "output_bytes": 511096312, | |
| "output_keys": 2062, | |
| "sample_reconstruction": { | |
| "down_proj": { | |
| "absolute_error_sum": 5.3247372668702155, | |
| "error_squared_sum": 0.008210588171982936, | |
| "max_absolute_error": 0.007254466414451599, | |
| "sample_values": 6144, | |
| "source_squared_sum": 0.9198178055735473 | |
| } | |
| }, | |
| "schema": "blackfrost.cyber_frost_nvfp4.shared_w13.v1", | |
| "shard_number": 47, | |
| "source": "model-00047-of-00131.safetensors", | |
| "storage_mode": "requantized_routed_experts_preserved_other_source_tensors" | |
| }, | |
| { | |
| "calibrated_gate_up_inputs_equal": true, | |
| "calibrated_input_scales_preserved_exactly": true, | |
| "elapsed_seconds": 10.667, | |
| "emitted_expert_matrices": 1024, | |
| "finite_positive_global_and_block_scales": true, | |
| "fused_source_tensors": 1, | |
| "gate_up_pairs_checked": 512, | |
| "output": "model-00048-of-00131.safetensors", | |
| "output_bytes": 944268824, | |
| "output_keys": 4096, | |
| "sample_reconstruction": { | |
| "gate_proj": { | |
| "absolute_error_sum": 5.637061126879416, | |
| "error_squared_sum": 0.00924638161879707, | |
| "max_absolute_error": 0.0073416568338871, | |
| "sample_values": 6144, | |
| "source_squared_sum": 1.0219392910245249 | |
| }, | |
| "up_proj": { | |
| "absolute_error_sum": 5.557981962570921, | |
| "error_squared_sum": 0.008833626753376261, | |
| "max_absolute_error": 0.006155829876661301, | |
| "sample_values": 6144, | |
| "source_squared_sum": 1.009090507660932 | |
| } | |
| }, | |
| "schema": "blackfrost.cyber_frost_nvfp4.shared_w13.v1", | |
| "shard_number": 48, | |
| "source": "model-00048-of-00131.safetensors", | |
| "storage_mode": "requantized_routed_experts_preserved_other_source_tensors" | |
| }, | |
| { | |
| "calibrated_gate_up_inputs_equal": true, | |
| "calibrated_input_scales_preserved_exactly": true, | |
| "elapsed_seconds": 4.492, | |
| "emitted_expert_matrices": 512, | |
| "finite_positive_global_and_block_scales": true, | |
| "fused_source_tensors": 1, | |
| "gate_up_pairs_checked": 0, | |
| "output": "model-00049-of-00131.safetensors", | |
| "output_bytes": 729848568, | |
| "output_keys": 2080, | |
| "sample_reconstruction": { | |
| "down_proj": { | |
| "absolute_error_sum": 5.34938176011201, | |
| "error_squared_sum": 0.008240497734392761, | |
| "max_absolute_error": 0.0058419369161129, | |
| "sample_values": 6144, | |
| "source_squared_sum": 0.9124136959936484 | |
| } | |
| }, | |
| "schema": "blackfrost.cyber_frost_nvfp4.shared_w13.v1", | |
| "shard_number": 49, | |
| "source": "model-00049-of-00131.safetensors", | |
| "storage_mode": "requantized_routed_experts_preserved_other_source_tensors" | |
| }, | |
| { | |
| "calibrated_gate_up_inputs_equal": true, | |
| "calibrated_input_scales_preserved_exactly": true, | |
| "elapsed_seconds": 11.16, | |
| "emitted_expert_matrices": 1024, | |
| "finite_positive_global_and_block_scales": true, | |
| "fused_source_tensors": 1, | |
| "gate_up_pairs_checked": 512, | |
| "output": "model-00050-of-00131.safetensors", | |
| "output_bytes": 944268824, | |
| "output_keys": 4096, | |
| "sample_reconstruction": { | |
| "gate_proj": { | |
| "absolute_error_sum": 5.538906019646674, | |
| "error_squared_sum": 0.009036899739076915, | |
| "max_absolute_error": 0.0073067806661129, | |
| "sample_values": 6144, | |
| "source_squared_sum": 0.9640389822737208 | |
| }, | |
| "up_proj": { | |
| "absolute_error_sum": 5.61992082814686, | |
| "error_squared_sum": 0.008866452706624513, | |
| "max_absolute_error": 0.006057010963559151, | |
| "sample_values": 6144, | |
| "source_squared_sum": 1.0123207087172101 | |
| } | |
| }, | |
| "schema": "blackfrost.cyber_frost_nvfp4.shared_w13.v1", | |
| "shard_number": 50, | |
| "source": "model-00050-of-00131.safetensors", | |
| "storage_mode": "requantized_routed_experts_preserved_other_source_tensors" | |
| }, | |
| { | |
| "calibrated_gate_up_inputs_equal": true, | |
| "calibrated_input_scales_preserved_exactly": true, | |
| "elapsed_seconds": 4.298, | |
| "emitted_expert_matrices": 512, | |
| "finite_positive_global_and_block_scales": true, | |
| "fused_source_tensors": 1, | |
| "gate_up_pairs_checked": 0, | |
| "output": "model-00051-of-00131.safetensors", | |
| "output_bytes": 626933712, | |
| "output_keys": 2070, | |
| "sample_reconstruction": { | |
| "down_proj": { | |
| "absolute_error_sum": 5.344488581176847, | |
| "error_squared_sum": 0.008357582197902232, | |
| "max_absolute_error": 0.006033763289451599, | |
| "sample_values": 6144, | |
| "source_squared_sum": 0.9414245546767983 | |
| } | |
| }, | |
| "schema": "blackfrost.cyber_frost_nvfp4.shared_w13.v1", | |
| "shard_number": 51, | |
| "source": "model-00051-of-00131.safetensors", | |
| "storage_mode": "requantized_routed_experts_preserved_other_source_tensors" | |
| }, | |
| { | |
| "calibrated_gate_up_inputs_equal": true, | |
| "calibrated_input_scales_preserved_exactly": true, | |
| "elapsed_seconds": 11.19, | |
| "emitted_expert_matrices": 1024, | |
| "finite_positive_global_and_block_scales": true, | |
| "fused_source_tensors": 1, | |
| "gate_up_pairs_checked": 512, | |
| "output": "model-00052-of-00131.safetensors", | |
| "output_bytes": 944268824, | |
| "output_keys": 4096, | |
| "sample_reconstruction": { | |
| "gate_proj": { | |
| "absolute_error_sum": 5.536034131422639, | |
| "error_squared_sum": 0.008994610863694772, | |
| "max_absolute_error": 0.0060686394572257996, | |
| "sample_values": 6144, | |
| "source_squared_sum": 0.9618893474033169 | |
| }, | |
| "up_proj": { | |
| "absolute_error_sum": 5.453077611629851, | |
| "error_squared_sum": 0.008581077739414746, | |
| "max_absolute_error": 0.005807060748338699, | |
| "sample_values": 6144, | |
| "source_squared_sum": 0.9578404597534309 | |
| } | |
| }, | |
| "schema": "blackfrost.cyber_frost_nvfp4.shared_w13.v1", | |
| "shard_number": 52, | |
| "source": "model-00052-of-00131.safetensors", | |
| "storage_mode": "requantized_routed_experts_preserved_other_source_tensors" | |
| }, | |
| { | |
| "calibrated_gate_up_inputs_equal": true, | |
| "calibrated_input_scales_preserved_exactly": true, | |
| "elapsed_seconds": 4.362, | |
| "emitted_expert_matrices": 512, | |
| "finite_positive_global_and_block_scales": true, | |
| "fused_source_tensors": 1, | |
| "gate_up_pairs_checked": 0, | |
| "output": "model-00053-of-00131.safetensors", | |
| "output_bytes": 633487456, | |
| "output_keys": 2071, | |
| "sample_reconstruction": { | |
| "down_proj": { | |
| "absolute_error_sum": 5.424936547293328, | |
| "error_squared_sum": 0.00841142888467136, | |
| "max_absolute_error": 0.00617326982319355, | |
| "sample_values": 6144, | |
| "source_squared_sum": 0.9393491528293794 | |
| } | |
| }, | |
| "schema": "blackfrost.cyber_frost_nvfp4.shared_w13.v1", | |
| "shard_number": 53, | |
| "source": "model-00053-of-00131.safetensors", | |
| "storage_mode": "requantized_routed_experts_preserved_other_source_tensors" | |
| }, | |
| { | |
| "calibrated_gate_up_inputs_equal": true, | |
| "calibrated_input_scales_preserved_exactly": true, | |
| "elapsed_seconds": 10.907, | |
| "emitted_expert_matrices": 1024, | |
| "finite_positive_global_and_block_scales": true, | |
| "fused_source_tensors": 1, | |
| "gate_up_pairs_checked": 512, | |
| "output": "model-00054-of-00131.safetensors", | |
| "output_bytes": 944268824, | |
| "output_keys": 4096, | |
| "sample_reconstruction": { | |
| "gate_proj": { | |
| "absolute_error_sum": 5.618128931266256, | |
| "error_squared_sum": 0.009153503537742856, | |
| "max_absolute_error": 0.006033763289451599, | |
| "sample_values": 6144, | |
| "source_squared_sum": 1.0223092524649395 | |
| }, | |
| "up_proj": { | |
| "absolute_error_sum": 5.623976738541387, | |
| "error_squared_sum": 0.009091236178055411, | |
| "max_absolute_error": 0.00582449696958065, | |
| "sample_values": 6144, | |
| "source_squared_sum": 1.0243618411870437 | |
| } | |
| }, | |
| "schema": "blackfrost.cyber_frost_nvfp4.shared_w13.v1", | |
| "shard_number": 54, | |
| "source": "model-00054-of-00131.safetensors", | |
| "storage_mode": "requantized_routed_experts_preserved_other_source_tensors" | |
| }, | |
| { | |
| "calibrated_gate_up_inputs_equal": true, | |
| "calibrated_input_scales_preserved_exactly": true, | |
| "elapsed_seconds": 4.021, | |
| "emitted_expert_matrices": 512, | |
| "finite_positive_global_and_block_scales": true, | |
| "fused_source_tensors": 1, | |
| "gate_up_pairs_checked": 0, | |
| "output": "model-00055-of-00131.safetensors", | |
| "output_bytes": 511014256, | |
| "output_keys": 2061, | |
| "sample_reconstruction": { | |
| "down_proj": { | |
| "absolute_error_sum": 5.509126932360232, | |
| "error_squared_sum": 0.008841632497527892, | |
| "max_absolute_error": 0.005859375, | |
| "sample_values": 6144, | |
| "source_squared_sum": 1.0024791096385184 | |
| } | |
| }, | |
| "schema": "blackfrost.cyber_frost_nvfp4.shared_w13.v1", | |
| "shard_number": 55, | |
| "source": "model-00055-of-00131.safetensors", | |
| "storage_mode": "requantized_routed_experts_preserved_other_source_tensors" | |
| }, | |
| { | |
| "calibrated_gate_up_inputs_equal": true, | |
| "calibrated_input_scales_preserved_exactly": true, | |
| "elapsed_seconds": 10.627, | |
| "emitted_expert_matrices": 1024, | |
| "finite_positive_global_and_block_scales": true, | |
| "fused_source_tensors": 1, | |
| "gate_up_pairs_checked": 512, | |
| "output": "model-00056-of-00131.safetensors", | |
| "output_bytes": 944268824, | |
| "output_keys": 4096, | |
| "sample_reconstruction": { | |
| "gate_proj": { | |
| "absolute_error_sum": 5.830275502055883, | |
| "error_squared_sum": 0.009870565238123703, | |
| "max_absolute_error": 0.0064871646463871, | |
| "sample_values": 6144, | |
| "source_squared_sum": 1.0794742752205053 | |
| }, | |
| "up_proj": { | |
| "absolute_error_sum": 5.796504138619639, | |
| "error_squared_sum": 0.009714901644970582, | |
| "max_absolute_error": 0.00626627542078495, | |
| "sample_values": 6144, | |
| "source_squared_sum": 1.0824622366258156 | |
| } | |
| }, | |
| "schema": "blackfrost.cyber_frost_nvfp4.shared_w13.v1", | |
| "shard_number": 56, | |
| "source": "model-00056-of-00131.safetensors", | |
| "storage_mode": "requantized_routed_experts_preserved_other_source_tensors" | |
| }, | |
| { | |
| "calibrated_gate_up_inputs_equal": true, | |
| "calibrated_input_scales_preserved_exactly": true, | |
| "elapsed_seconds": 4.522, | |
| "emitted_expert_matrices": 512, | |
| "finite_positive_global_and_block_scales": true, | |
| "fused_source_tensors": 1, | |
| "gate_up_pairs_checked": 0, | |
| "output": "model-00057-of-00131.safetensors", | |
| "output_bytes": 736381696, | |
| "output_keys": 2080, | |
| "sample_reconstruction": { | |
| "down_proj": { | |
| "absolute_error_sum": 5.5811587981297635, | |
| "error_squared_sum": 0.009182986528995778, | |
| "max_absolute_error": 0.006173271685838699, | |
| "sample_values": 6144, | |
| "source_squared_sum": 0.9918365396716544 | |
| } | |
| }, | |
| "schema": "blackfrost.cyber_frost_nvfp4.shared_w13.v1", | |
| "shard_number": 57, | |
| "source": "model-00057-of-00131.safetensors", | |
| "storage_mode": "requantized_routed_experts_preserved_other_source_tensors" | |
| }, | |
| { | |
| "calibrated_gate_up_inputs_equal": true, | |
| "calibrated_input_scales_preserved_exactly": true, | |
| "elapsed_seconds": 11.694, | |
| "emitted_expert_matrices": 1024, | |
| "finite_positive_global_and_block_scales": true, | |
| "fused_source_tensors": 1, | |
| "gate_up_pairs_checked": 512, | |
| "output": "model-00058-of-00131.safetensors", | |
| "output_bytes": 944264728, | |
| "output_keys": 4096, | |
| "sample_reconstruction": { | |
| "gate_proj": { | |
| "absolute_error_sum": 5.524326032784302, | |
| "error_squared_sum": 0.008803107100121589, | |
| "max_absolute_error": 0.00558035634458065, | |
| "sample_values": 6144, | |
| "source_squared_sum": 0.9850473465814424 | |
| }, | |
| "up_proj": { | |
| "absolute_error_sum": 5.668600690085441, | |
| "error_squared_sum": 0.009295016652061611, | |
| "max_absolute_error": 0.007707871496677399, | |
| "sample_values": 6144, | |
| "source_squared_sum": 1.033483811380641 | |
| } | |
| }, | |
| "schema": "blackfrost.cyber_frost_nvfp4.shared_w13.v1", | |
| "shard_number": 58, | |
| "source": "model-00058-of-00131.safetensors", | |
| "storage_mode": "requantized_routed_experts_preserved_other_source_tensors" | |
| }, | |
| { | |
| "calibrated_gate_up_inputs_equal": true, | |
| "calibrated_input_scales_preserved_exactly": true, | |
| "elapsed_seconds": 4.172, | |
| "emitted_expert_matrices": 512, | |
| "finite_positive_global_and_block_scales": true, | |
| "fused_source_tensors": 1, | |
| "gate_up_pairs_checked": 0, | |
| "output": "model-00059-of-00131.safetensors", | |
| "output_bytes": 511012416, | |
| "output_keys": 2062, | |
| "sample_reconstruction": { | |
| "down_proj": { | |
| "absolute_error_sum": 5.503483554231934, | |
| "error_squared_sum": 0.00865161627674566, | |
| "max_absolute_error": 0.006644114851951599, | |
| "sample_values": 6144, | |
| "source_squared_sum": 0.9561847652348955 | |
| } | |
| }, | |
| "schema": "blackfrost.cyber_frost_nvfp4.shared_w13.v1", | |
| "shard_number": 59, | |
| "source": "model-00059-of-00131.safetensors", | |
| "storage_mode": "requantized_routed_experts_preserved_other_source_tensors" | |
| }, | |
| { | |
| "calibrated_gate_up_inputs_equal": true, | |
| "calibrated_input_scales_preserved_exactly": true, | |
| "elapsed_seconds": 6.537, | |
| "emitted_expert_matrices": 0, | |
| "finite_positive_global_and_block_scales": true, | |
| "fused_source_tensors": 0, | |
| "gate_up_pairs_checked": 0, | |
| "output": "model-00060-of-00131.safetensors", | |
| "output_bytes": 3471361528, | |
| "output_keys": 9, | |
| "sample_reconstruction": {}, | |
| "schema": "blackfrost.cyber_frost_nvfp4.shared_w13.v1", | |
| "shard_number": 60, | |
| "source": "model-00060-of-00131.safetensors", | |
| "storage_mode": "copy_unchanged_cross_device" | |
| }, | |
| { | |
| "calibrated_gate_up_inputs_equal": true, | |
| "calibrated_input_scales_preserved_exactly": true, | |
| "elapsed_seconds": 11.232, | |
| "emitted_expert_matrices": 1024, | |
| "finite_positive_global_and_block_scales": true, | |
| "fused_source_tensors": 1, | |
| "gate_up_pairs_checked": 512, | |
| "output": "model-00061-of-00131.safetensors", | |
| "output_bytes": 944268824, | |
| "output_keys": 4096, | |
| "sample_reconstruction": { | |
| "gate_proj": { | |
| "absolute_error_sum": 5.522524136584252, | |
| "error_squared_sum": 0.008703740491741375, | |
| "max_absolute_error": 0.005859373137354851, | |
| "sample_values": 6144, | |
| "source_squared_sum": 0.9741713342084859 | |
| }, | |
| "up_proj": { | |
| "absolute_error_sum": 5.460364357451908, | |
| "error_squared_sum": 0.008799773784802272, | |
| "max_absolute_error": 0.0064290352165699005, | |
| "sample_values": 6144, | |
| "source_squared_sum": 0.9581500086691062 | |
| } | |
| }, | |
| "schema": "blackfrost.cyber_frost_nvfp4.shared_w13.v1", | |
| "shard_number": 61, | |
| "source": "model-00061-of-00131.safetensors", | |
| "storage_mode": "requantized_routed_experts_preserved_other_source_tensors" | |
| }, | |
| { | |
| "calibrated_gate_up_inputs_equal": true, | |
| "calibrated_input_scales_preserved_exactly": true, | |
| "elapsed_seconds": 5.859, | |
| "emitted_expert_matrices": 512, | |
| "finite_positive_global_and_block_scales": true, | |
| "fused_source_tensors": 1, | |
| "gate_up_pairs_checked": 0, | |
| "output": "model-00062-of-00131.safetensors", | |
| "output_bytes": 2304657472, | |
| "output_keys": 2071, | |
| "sample_reconstruction": { | |
| "down_proj": { | |
| "absolute_error_sum": 5.387690067174844, | |
| "error_squared_sum": 0.008412880609429319, | |
| "max_absolute_error": 0.0074637271463871, | |
| "sample_values": 6144, | |
| "source_squared_sum": 0.903125052338777 | |
| } | |
| }, | |
| "schema": "blackfrost.cyber_frost_nvfp4.shared_w13.v1", | |
| "shard_number": 62, | |
| "source": "model-00062-of-00131.safetensors", | |
| "storage_mode": "requantized_routed_experts_preserved_other_source_tensors" | |
| }, | |
| { | |
| "calibrated_gate_up_inputs_equal": true, | |
| "calibrated_input_scales_preserved_exactly": true, | |
| "elapsed_seconds": 10.831, | |
| "emitted_expert_matrices": 1024, | |
| "finite_positive_global_and_block_scales": true, | |
| "fused_source_tensors": 1, | |
| "gate_up_pairs_checked": 512, | |
| "output": "model-00063-of-00131.safetensors", | |
| "output_bytes": 944268824, | |
| "output_keys": 4096, | |
| "sample_reconstruction": { | |
| "gate_proj": { | |
| "absolute_error_sum": 5.655583991319872, | |
| "error_squared_sum": 0.009060862116599101, | |
| "max_absolute_error": 0.006103515625, | |
| "sample_values": 6144, | |
| "source_squared_sum": 1.0045363851508438 | |
| }, | |
| "up_proj": { | |
| "absolute_error_sum": 5.484193931682967, | |
| "error_squared_sum": 0.008487697674711938, | |
| "max_absolute_error": 0.005719862878322601, | |
| "sample_values": 6144, | |
| "source_squared_sum": 0.9570350598238886 | |
| } | |
| }, | |
| "schema": "blackfrost.cyber_frost_nvfp4.shared_w13.v1", | |
| "shard_number": 63, | |
| "source": "model-00063-of-00131.safetensors", | |
| "storage_mode": "requantized_routed_experts_preserved_other_source_tensors" | |
| }, | |
| { | |
| "calibrated_gate_up_inputs_equal": true, | |
| "calibrated_input_scales_preserved_exactly": true, | |
| "elapsed_seconds": 4.439, | |
| "emitted_expert_matrices": 512, | |
| "finite_positive_global_and_block_scales": true, | |
| "fused_source_tensors": 1, | |
| "gate_up_pairs_checked": 0, | |
| "output": "model-00064-of-00131.safetensors", | |
| "output_bytes": 626933712, | |
| "output_keys": 2070, | |
| "sample_reconstruction": { | |
| "down_proj": { | |
| "absolute_error_sum": 5.394355252501555, | |
| "error_squared_sum": 0.008356470356096175, | |
| "max_absolute_error": 0.0062255859375, | |
| "sample_values": 6144, | |
| "source_squared_sum": 0.9289657916609839 | |
| } | |
| }, | |
| "schema": "blackfrost.cyber_frost_nvfp4.shared_w13.v1", | |
| "shard_number": 64, | |
| "source": "model-00064-of-00131.safetensors", | |
| "storage_mode": "requantized_routed_experts_preserved_other_source_tensors" | |
| }, | |
| { | |
| "calibrated_gate_up_inputs_equal": true, | |
| "calibrated_input_scales_preserved_exactly": true, | |
| "elapsed_seconds": 11.244, | |
| "emitted_expert_matrices": 1024, | |
| "finite_positive_global_and_block_scales": true, | |
| "fused_source_tensors": 1, | |
| "gate_up_pairs_checked": 512, | |
| "output": "model-00065-of-00131.safetensors", | |
| "output_bytes": 944268824, | |
| "output_keys": 4096, | |
| "sample_reconstruction": { | |
| "gate_proj": { | |
| "absolute_error_sum": 5.703900416614488, | |
| "error_squared_sum": 0.009386179083405504, | |
| "max_absolute_error": 0.006789432838559151, | |
| "sample_values": 6144, | |
| "source_squared_sum": 1.0381944529735876 | |
| }, | |
| "up_proj": { | |
| "absolute_error_sum": 5.41939201799687, | |
| "error_squared_sum": 0.008604839209421867, | |
| "max_absolute_error": 0.005789622664451599, | |
| "sample_values": 6144, | |
| "source_squared_sum": 0.9358243875354013 | |
| } | |
| }, | |
| "schema": "blackfrost.cyber_frost_nvfp4.shared_w13.v1", | |
| "shard_number": 65, | |
| "source": "model-00065-of-00131.safetensors", | |
| "storage_mode": "requantized_routed_experts_preserved_other_source_tensors" | |
| }, | |
| { | |
| "calibrated_gate_up_inputs_equal": true, | |
| "calibrated_input_scales_preserved_exactly": true, | |
| "elapsed_seconds": 4.379, | |
| "emitted_expert_matrices": 512, | |
| "finite_positive_global_and_block_scales": true, | |
| "fused_source_tensors": 1, | |
| "gate_up_pairs_checked": 0, | |
| "output": "model-00066-of-00131.safetensors", | |
| "output_bytes": 513635808, | |
| "output_keys": 2062, | |
| "sample_reconstruction": { | |
| "down_proj": { | |
| "absolute_error_sum": 5.279753647046164, | |
| "error_squared_sum": 0.008074495039287685, | |
| "max_absolute_error": 0.0053420290350914, | |
| "sample_values": 6144, | |
| "source_squared_sum": 0.901296195958134 | |
| } | |
| }, | |
| "schema": "blackfrost.cyber_frost_nvfp4.shared_w13.v1", | |
| "shard_number": 66, | |
| "source": "model-00066-of-00131.safetensors", | |
| "storage_mode": "requantized_routed_experts_preserved_other_source_tensors" | |
| }, | |
| { | |
| "calibrated_gate_up_inputs_equal": true, | |
| "calibrated_input_scales_preserved_exactly": true, | |
| "elapsed_seconds": 11.262, | |
| "emitted_expert_matrices": 1024, | |
| "finite_positive_global_and_block_scales": true, | |
| "fused_source_tensors": 1, | |
| "gate_up_pairs_checked": 512, | |
| "output": "model-00067-of-00131.safetensors", | |
| "output_bytes": 944268824, | |
| "output_keys": 4096, | |
| "sample_reconstruction": { | |
| "gate_proj": { | |
| "absolute_error_sum": 5.385505047044717, | |
| "error_squared_sum": 0.00824238845591854, | |
| "max_absolute_error": 0.0056036077439785, | |
| "sample_values": 6144, | |
| "source_squared_sum": 0.9587984130361642 | |
| }, | |
| "up_proj": { | |
| "absolute_error_sum": 5.544245528639294, | |
| "error_squared_sum": 0.008949520551008594, | |
| "max_absolute_error": 0.0072195865213871, | |
| "sample_values": 6144, | |
| "source_squared_sum": 1.020594535550492 | |
| } | |
| }, | |
| "schema": "blackfrost.cyber_frost_nvfp4.shared_w13.v1", | |
| "shard_number": 67, | |
| "source": "model-00067-of-00131.safetensors", | |
| "storage_mode": "requantized_routed_experts_preserved_other_source_tensors" | |
| }, | |
| { | |
| "calibrated_gate_up_inputs_equal": true, | |
| "calibrated_input_scales_preserved_exactly": true, | |
| "elapsed_seconds": 4.52, | |
| "emitted_expert_matrices": 512, | |
| "finite_positive_global_and_block_scales": true, | |
| "fused_source_tensors": 1, | |
| "gate_up_pairs_checked": 0, | |
| "output": "model-00068-of-00131.safetensors", | |
| "output_bytes": 729827968, | |
| "output_keys": 2079, | |
| "sample_reconstruction": { | |
| "down_proj": { | |
| "absolute_error_sum": 5.515024423657451, | |
| "error_squared_sum": 0.008942022928950638, | |
| "max_absolute_error": 0.005789618939161301, | |
| "sample_values": 6144, | |
| "source_squared_sum": 0.9679185765707246 | |
| } | |
| }, | |
| "schema": "blackfrost.cyber_frost_nvfp4.shared_w13.v1", | |
| "shard_number": 68, | |
| "source": "model-00068-of-00131.safetensors", | |
| "storage_mode": "requantized_routed_experts_preserved_other_source_tensors" | |
| }, | |
| { | |
| "calibrated_gate_up_inputs_equal": true, | |
| "calibrated_input_scales_preserved_exactly": true, | |
| "elapsed_seconds": 10.867, | |
| "emitted_expert_matrices": 1024, | |
| "finite_positive_global_and_block_scales": true, | |
| "fused_source_tensors": 1, | |
| "gate_up_pairs_checked": 512, | |
| "output": "model-00069-of-00131.safetensors", | |
| "output_bytes": 944268824, | |
| "output_keys": 4096, | |
| "sample_reconstruction": { | |
| "gate_proj": { | |
| "absolute_error_sum": 5.606244854396209, | |
| "error_squared_sum": 0.009151318223298835, | |
| "max_absolute_error": 0.006033763289451599, | |
| "sample_values": 6144, | |
| "source_squared_sum": 0.9920629746355362 | |
| }, | |
| "up_proj": { | |
| "absolute_error_sum": 5.4834565030178055, | |
| "error_squared_sum": 0.008646159559835699, | |
| "max_absolute_error": 0.007393971085548401, | |
| "sample_values": 6144, | |
| "source_squared_sum": 0.9722557122739213 | |
| } | |
| }, | |
| "schema": "blackfrost.cyber_frost_nvfp4.shared_w13.v1", | |
| "shard_number": 69, | |
| "source": "model-00069-of-00131.safetensors", | |
| "storage_mode": "requantized_routed_experts_preserved_other_source_tensors" | |
| }, | |
| { | |
| "calibrated_gate_up_inputs_equal": true, | |
| "calibrated_input_scales_preserved_exactly": true, | |
| "elapsed_seconds": 4.136, | |
| "emitted_expert_matrices": 512, | |
| "finite_positive_global_and_block_scales": true, | |
| "fused_source_tensors": 1, | |
| "gate_up_pairs_checked": 0, | |
| "output": "model-00070-of-00131.safetensors", | |
| "output_bytes": 626938952, | |
| "output_keys": 2071, | |
| "sample_reconstruction": { | |
| "down_proj": { | |
| "absolute_error_sum": 5.495457615004852, | |
| "error_squared_sum": 0.008602351209895982, | |
| "max_absolute_error": 0.0058942511677742004, | |
| "sample_values": 6144, | |
| "source_squared_sum": 0.9677656315263419 | |
| } | |
| }, | |
| "schema": "blackfrost.cyber_frost_nvfp4.shared_w13.v1", | |
| "shard_number": 70, | |
| "source": "model-00070-of-00131.safetensors", | |
| "storage_mode": "requantized_routed_experts_preserved_other_source_tensors" | |
| }, | |
| { | |
| "calibrated_gate_up_inputs_equal": true, | |
| "calibrated_input_scales_preserved_exactly": true, | |
| "elapsed_seconds": 11.205, | |
| "emitted_expert_matrices": 1024, | |
| "finite_positive_global_and_block_scales": true, | |
| "fused_source_tensors": 1, | |
| "gate_up_pairs_checked": 512, | |
| "output": "model-00071-of-00131.safetensors", | |
| "output_bytes": 944268824, | |
| "output_keys": 4096, | |
| "sample_reconstruction": { | |
| "gate_proj": { | |
| "absolute_error_sum": 5.499343150295317, | |
| "error_squared_sum": 0.008799860029693364, | |
| "max_absolute_error": 0.0066266730427742004, | |
| "sample_values": 6144, | |
| "source_squared_sum": 0.9642303399749066 | |
| }, | |
| "up_proj": { | |
| "absolute_error_sum": 5.460399418021552, | |
| "error_squared_sum": 0.00853239989965486, | |
| "max_absolute_error": 0.006545292213559151, | |
| "sample_values": 6144, | |
| "source_squared_sum": 0.9754617291510428 | |
| } | |
| }, | |
| "schema": "blackfrost.cyber_frost_nvfp4.shared_w13.v1", | |
| "shard_number": 71, | |
| "source": "model-00071-of-00131.safetensors", | |
| "storage_mode": "requantized_routed_experts_preserved_other_source_tensors" | |
| }, | |
| { | |
| "calibrated_gate_up_inputs_equal": true, | |
| "calibrated_input_scales_preserved_exactly": true, | |
| "elapsed_seconds": 4.292, | |
| "emitted_expert_matrices": 512, | |
| "finite_positive_global_and_block_scales": true, | |
| "fused_source_tensors": 1, | |
| "gate_up_pairs_checked": 0, | |
| "output": "model-00072-of-00131.safetensors", | |
| "output_bytes": 633487560, | |
| "output_keys": 2072, | |
| "sample_reconstruction": { | |
| "down_proj": { | |
| "absolute_error_sum": 5.535992015618831, | |
| "error_squared_sum": 0.008780621029366404, | |
| "max_absolute_error": 0.00619070790708065, | |
| "sample_values": 6144, | |
| "source_squared_sum": 0.9812492806059345 | |
| } | |
| }, | |
| "schema": "blackfrost.cyber_frost_nvfp4.shared_w13.v1", | |
| "shard_number": 72, | |
| "source": "model-00072-of-00131.safetensors", | |
| "storage_mode": "requantized_routed_experts_preserved_other_source_tensors" | |
| }, | |
| { | |
| "calibrated_gate_up_inputs_equal": true, | |
| "calibrated_input_scales_preserved_exactly": true, | |
| "elapsed_seconds": 11.748, | |
| "emitted_expert_matrices": 1024, | |
| "finite_positive_global_and_block_scales": true, | |
| "fused_source_tensors": 1, | |
| "gate_up_pairs_checked": 512, | |
| "output": "model-00073-of-00131.safetensors", | |
| "output_bytes": 944268824, | |
| "output_keys": 4096, | |
| "sample_reconstruction": { | |
| "gate_proj": { | |
| "absolute_error_sum": 5.34134764468763, | |
| "error_squared_sum": 0.008238978866872422, | |
| "max_absolute_error": 0.00613839365541935, | |
| "sample_values": 6144, | |
| "source_squared_sum": 0.93500419710532 | |
| }, | |
| "up_proj": { | |
| "absolute_error_sum": 5.511580316931941, | |
| "error_squared_sum": 0.00881529684844274, | |
| "max_absolute_error": 0.00638253428041935, | |
| "sample_values": 6144, | |
| "source_squared_sum": 0.986355260339951 | |
| } | |
| }, | |
| "schema": "blackfrost.cyber_frost_nvfp4.shared_w13.v1", | |
| "shard_number": 73, | |
| "source": "model-00073-of-00131.safetensors", | |
| "storage_mode": "requantized_routed_experts_preserved_other_source_tensors" | |
| }, | |
| { | |
| "calibrated_gate_up_inputs_equal": true, | |
| "calibrated_input_scales_preserved_exactly": true, | |
| "elapsed_seconds": 4.105, | |
| "emitted_expert_matrices": 512, | |
| "finite_positive_global_and_block_scales": true, | |
| "fused_source_tensors": 1, | |
| "gate_up_pairs_checked": 0, | |
| "output": "model-00074-of-00131.safetensors", | |
| "output_bytes": 511014256, | |
| "output_keys": 2061, | |
| "sample_reconstruction": { | |
| "down_proj": { | |
| "absolute_error_sum": 5.5504602221772075, | |
| "error_squared_sum": 0.008957867336787498, | |
| "max_absolute_error": 0.00634765625, | |
| "sample_values": 6144, | |
| "source_squared_sum": 0.9964283625540455 | |
| } | |
| }, | |
| "schema": "blackfrost.cyber_frost_nvfp4.shared_w13.v1", | |
| "shard_number": 74, | |
| "source": "model-00074-of-00131.safetensors", | |
| "storage_mode": "requantized_routed_experts_preserved_other_source_tensors" | |
| }, | |
| { | |
| "calibrated_gate_up_inputs_equal": true, | |
| "calibrated_input_scales_preserved_exactly": true, | |
| "elapsed_seconds": 11.564, | |
| "emitted_expert_matrices": 1024, | |
| "finite_positive_global_and_block_scales": true, | |
| "fused_source_tensors": 1, | |
| "gate_up_pairs_checked": 512, | |
| "output": "model-00075-of-00131.safetensors", | |
| "output_bytes": 944268824, | |
| "output_keys": 4096, | |
| "sample_reconstruction": { | |
| "gate_proj": { | |
| "absolute_error_sum": 5.346434296690859, | |
| "error_squared_sum": 0.00815278328929377, | |
| "max_absolute_error": 0.0058244988322257996, | |
| "sample_values": 6144, | |
| "source_squared_sum": 0.9036091479843797 | |
| }, | |
| "up_proj": { | |
| "absolute_error_sum": 5.505349541665055, | |
| "error_squared_sum": 0.008825055707901102, | |
| "max_absolute_error": 0.006661549210548401, | |
| "sample_values": 6144, | |
| "source_squared_sum": 0.9732888324801097 | |
| } | |
| }, | |
| "schema": "blackfrost.cyber_frost_nvfp4.shared_w13.v1", | |
| "shard_number": 75, | |
| "source": "model-00075-of-00131.safetensors", | |
| "storage_mode": "requantized_routed_experts_preserved_other_source_tensors" | |
| }, | |
| { | |
| "calibrated_gate_up_inputs_equal": true, | |
| "calibrated_input_scales_preserved_exactly": true, | |
| "elapsed_seconds": 4.696, | |
| "emitted_expert_matrices": 512, | |
| "finite_positive_global_and_block_scales": true, | |
| "fused_source_tensors": 1, | |
| "gate_up_pairs_checked": 0, | |
| "output": "model-00076-of-00131.safetensors", | |
| "output_bytes": 733104888, | |
| "output_keys": 2080, | |
| "sample_reconstruction": { | |
| "down_proj": { | |
| "absolute_error_sum": 5.607156103476882, | |
| "error_squared_sum": 0.00917554530784145, | |
| "max_absolute_error": 0.00603376142680645, | |
| "sample_values": 6144, | |
| "source_squared_sum": 1.006071647872914 | |
| } | |
| }, | |
| "schema": "blackfrost.cyber_frost_nvfp4.shared_w13.v1", | |
| "shard_number": 76, | |
| "source": "model-00076-of-00131.safetensors", | |
| "storage_mode": "requantized_routed_experts_preserved_other_source_tensors" | |
| }, | |
| { | |
| "calibrated_gate_up_inputs_equal": true, | |
| "calibrated_input_scales_preserved_exactly": true, | |
| "elapsed_seconds": 10.735, | |
| "emitted_expert_matrices": 1024, | |
| "finite_positive_global_and_block_scales": true, | |
| "fused_source_tensors": 1, | |
| "gate_up_pairs_checked": 512, | |
| "output": "model-00077-of-00131.safetensors", | |
| "output_bytes": 944268824, | |
| "output_keys": 4096, | |
| "sample_reconstruction": { | |
| "gate_proj": { | |
| "absolute_error_sum": 5.549647584790364, | |
| "error_squared_sum": 0.008801789823617453, | |
| "max_absolute_error": 0.006394157186150551, | |
| "sample_values": 6144, | |
| "source_squared_sum": 0.9983981254139462 | |
| }, | |
| "up_proj": { | |
| "absolute_error_sum": 5.715491489274427, | |
| "error_squared_sum": 0.009618918604403523, | |
| "max_absolute_error": 0.0077311210334300995, | |
| "sample_values": 6144, | |
| "source_squared_sum": 1.0292322414230592 | |
| } | |
| }, | |
| "schema": "blackfrost.cyber_frost_nvfp4.shared_w13.v1", | |
| "shard_number": 77, | |
| "source": "model-00077-of-00131.safetensors", | |
| "storage_mode": "requantized_routed_experts_preserved_other_source_tensors" | |
| }, | |
| { | |
| "calibrated_gate_up_inputs_equal": true, | |
| "calibrated_input_scales_preserved_exactly": true, | |
| "elapsed_seconds": 4.253, | |
| "emitted_expert_matrices": 512, | |
| "finite_positive_global_and_block_scales": true, | |
| "fused_source_tensors": 1, | |
| "gate_up_pairs_checked": 0, | |
| "output": "model-00078-of-00131.safetensors", | |
| "output_bytes": 627015768, | |
| "output_keys": 2071, | |
| "sample_reconstruction": { | |
| "down_proj": { | |
| "absolute_error_sum": 5.62926771666389, | |
| "error_squared_sum": 0.00917606561753016, | |
| "max_absolute_error": 0.0057198647409677505, | |
| "sample_values": 6144, | |
| "source_squared_sum": 1.024618221842811 | |
| } | |
| }, | |
| "schema": "blackfrost.cyber_frost_nvfp4.shared_w13.v1", | |
| "shard_number": 78, | |
| "source": "model-00078-of-00131.safetensors", | |
| "storage_mode": "requantized_routed_experts_preserved_other_source_tensors" | |
| }, | |
| { | |
| "calibrated_gate_up_inputs_equal": true, | |
| "calibrated_input_scales_preserved_exactly": true, | |
| "elapsed_seconds": 10.734, | |
| "emitted_expert_matrices": 1024, | |
| "finite_positive_global_and_block_scales": true, | |
| "fused_source_tensors": 1, | |
| "gate_up_pairs_checked": 512, | |
| "output": "model-00079-of-00131.safetensors", | |
| "output_bytes": 944268824, | |
| "output_keys": 4096, | |
| "sample_reconstruction": { | |
| "gate_proj": { | |
| "absolute_error_sum": 5.553870598669164, | |
| "error_squared_sum": 0.008899605802615134, | |
| "max_absolute_error": 0.005545482039451599, | |
| "sample_values": 6144, | |
| "source_squared_sum": 0.9778951390235879 | |
| }, | |
| "up_proj": { | |
| "absolute_error_sum": 5.702822175808251, | |
| "error_squared_sum": 0.009540113581686747, | |
| "max_absolute_error": 0.0071149542927742004, | |
| "sample_values": 6144, | |
| "source_squared_sum": 1.0202292403964501 | |
| } | |
| }, | |
| "schema": "blackfrost.cyber_frost_nvfp4.shared_w13.v1", | |
| "shard_number": 79, | |
| "source": "model-00079-of-00131.safetensors", | |
| "storage_mode": "requantized_routed_experts_preserved_other_source_tensors" | |
| }, | |
| { | |
| "calibrated_gate_up_inputs_equal": true, | |
| "calibrated_input_scales_preserved_exactly": true, | |
| "elapsed_seconds": 4.07, | |
| "emitted_expert_matrices": 512, | |
| "finite_positive_global_and_block_scales": true, | |
| "fused_source_tensors": 1, | |
| "gate_up_pairs_checked": 0, | |
| "output": "model-00080-of-00131.safetensors", | |
| "output_bytes": 511014256, | |
| "output_keys": 2061, | |
| "sample_reconstruction": { | |
| "down_proj": { | |
| "absolute_error_sum": 5.585813673911616, | |
| "error_squared_sum": 0.008900968847862109, | |
| "max_absolute_error": 0.00602213479578495, | |
| "sample_values": 6144, | |
| "source_squared_sum": 1.0099520049108164 | |
| } | |
| }, | |
| "schema": "blackfrost.cyber_frost_nvfp4.shared_w13.v1", | |
| "shard_number": 80, | |
| "source": "model-00080-of-00131.safetensors", | |
| "storage_mode": "requantized_routed_experts_preserved_other_source_tensors" | |
| }, | |
| { | |
| "calibrated_gate_up_inputs_equal": true, | |
| "calibrated_input_scales_preserved_exactly": true, | |
| "elapsed_seconds": 11.213, | |
| "emitted_expert_matrices": 1024, | |
| "finite_positive_global_and_block_scales": true, | |
| "fused_source_tensors": 1, | |
| "gate_up_pairs_checked": 512, | |
| "output": "model-00081-of-00131.safetensors", | |
| "output_bytes": 944264728, | |
| "output_keys": 4096, | |
| "sample_reconstruction": { | |
| "gate_proj": { | |
| "absolute_error_sum": 5.574324888235424, | |
| "error_squared_sum": 0.009086207763053602, | |
| "max_absolute_error": 0.00732421875, | |
| "sample_values": 6144, | |
| "source_squared_sum": 1.0386743080658374 | |
| }, | |
| "up_proj": { | |
| "absolute_error_sum": 5.584553265012801, | |
| "error_squared_sum": 0.009073332663334417, | |
| "max_absolute_error": 0.0062081478536129, | |
| "sample_values": 6144, | |
| "source_squared_sum": 1.0310582872774718 | |
| } | |
| }, | |
| "schema": "blackfrost.cyber_frost_nvfp4.shared_w13.v1", | |
| "shard_number": 81, | |
| "source": "model-00081-of-00131.safetensors", | |
| "storage_mode": "requantized_routed_experts_preserved_other_source_tensors" | |
| }, | |
| { | |
| "calibrated_gate_up_inputs_equal": true, | |
| "calibrated_input_scales_preserved_exactly": true, | |
| "elapsed_seconds": 4.359, | |
| "emitted_expert_matrices": 512, | |
| "finite_positive_global_and_block_scales": true, | |
| "fused_source_tensors": 1, | |
| "gate_up_pairs_checked": 0, | |
| "output": "model-00082-of-00131.safetensors", | |
| "output_bytes": 729846504, | |
| "output_keys": 2080, | |
| "sample_reconstruction": { | |
| "down_proj": { | |
| "absolute_error_sum": 5.2783462312072515, | |
| "error_squared_sum": 0.008073420271795083, | |
| "max_absolute_error": 0.0052838996052742004, | |
| "sample_values": 6144, | |
| "source_squared_sum": 0.8847001052052226 | |
| } | |
| }, | |
| "schema": "blackfrost.cyber_frost_nvfp4.shared_w13.v1", | |
| "shard_number": 82, | |
| "source": "model-00082-of-00131.safetensors", | |
| "storage_mode": "requantized_routed_experts_preserved_other_source_tensors" | |
| }, | |
| { | |
| "calibrated_gate_up_inputs_equal": true, | |
| "calibrated_input_scales_preserved_exactly": true, | |
| "elapsed_seconds": 11.366, | |
| "emitted_expert_matrices": 1024, | |
| "finite_positive_global_and_block_scales": true, | |
| "fused_source_tensors": 1, | |
| "gate_up_pairs_checked": 512, | |
| "output": "model-00083-of-00131.safetensors", | |
| "output_bytes": 944268824, | |
| "output_keys": 4096, | |
| "sample_reconstruction": { | |
| "gate_proj": { | |
| "absolute_error_sum": 5.413586184149608, | |
| "error_squared_sum": 0.00843775257931753, | |
| "max_absolute_error": 0.005510598421096802, | |
| "sample_values": 6144, | |
| "source_squared_sum": 0.9465823282195824 | |
| }, | |
| "up_proj": { | |
| "absolute_error_sum": 5.51532444357872, | |
| "error_squared_sum": 0.008771207785240633, | |
| "max_absolute_error": 0.00687081553041935, | |
| "sample_values": 6144, | |
| "source_squared_sum": 0.9855831621412738 | |
| } | |
| }, | |
| "schema": "blackfrost.cyber_frost_nvfp4.shared_w13.v1", | |
| "shard_number": 83, | |
| "source": "model-00083-of-00131.safetensors", | |
| "storage_mode": "requantized_routed_experts_preserved_other_source_tensors" | |
| }, | |
| { | |
| "calibrated_gate_up_inputs_equal": true, | |
| "calibrated_input_scales_preserved_exactly": true, | |
| "elapsed_seconds": 4.403, | |
| "emitted_expert_matrices": 512, | |
| "finite_positive_global_and_block_scales": true, | |
| "fused_source_tensors": 1, | |
| "gate_up_pairs_checked": 0, | |
| "output": "model-00084-of-00131.safetensors", | |
| "output_bytes": 511014256, | |
| "output_keys": 2061, | |
| "sample_reconstruction": { | |
| "down_proj": { | |
| "absolute_error_sum": 5.586027930141427, | |
| "error_squared_sum": 0.009029707155614797, | |
| "max_absolute_error": 0.006574362516403198, | |
| "sample_values": 6144, | |
| "source_squared_sum": 1.0085909216236877 | |
| } | |
| }, | |
| "schema": "blackfrost.cyber_frost_nvfp4.shared_w13.v1", | |
| "shard_number": 84, | |
| "source": "model-00084-of-00131.safetensors", | |
| "storage_mode": "requantized_routed_experts_preserved_other_source_tensors" | |
| }, | |
| { | |
| "calibrated_gate_up_inputs_equal": true, | |
| "calibrated_input_scales_preserved_exactly": true, | |
| "elapsed_seconds": 10.521, | |
| "emitted_expert_matrices": 1024, | |
| "finite_positive_global_and_block_scales": true, | |
| "fused_source_tensors": 1, | |
| "gate_up_pairs_checked": 512, | |
| "output": "model-00085-of-00131.safetensors", | |
| "output_bytes": 944268824, | |
| "output_keys": 4096, | |
| "sample_reconstruction": { | |
| "gate_proj": { | |
| "absolute_error_sum": 5.759938788483851, | |
| "error_squared_sum": 0.009572849218906302, | |
| "max_absolute_error": 0.007010325789451599, | |
| "sample_values": 6144, | |
| "source_squared_sum": 1.0457971975826947 | |
| }, | |
| "up_proj": { | |
| "absolute_error_sum": 5.7992438898654655, | |
| "error_squared_sum": 0.00980984027626893, | |
| "max_absolute_error": 0.006905693560838699, | |
| "sample_values": 6144, | |
| "source_squared_sum": 1.0577050765832292 | |
| } | |
| }, | |
| "schema": "blackfrost.cyber_frost_nvfp4.shared_w13.v1", | |
| "shard_number": 85, | |
| "source": "model-00085-of-00131.safetensors", | |
| "storage_mode": "requantized_routed_experts_preserved_other_source_tensors" | |
| }, | |
| { | |
| "calibrated_gate_up_inputs_equal": true, | |
| "calibrated_input_scales_preserved_exactly": true, | |
| "elapsed_seconds": 4.32, | |
| "emitted_expert_matrices": 512, | |
| "finite_positive_global_and_block_scales": true, | |
| "fused_source_tensors": 1, | |
| "gate_up_pairs_checked": 0, | |
| "output": "model-00086-of-00131.safetensors", | |
| "output_bytes": 736381704, | |
| "output_keys": 2080, | |
| "sample_reconstruction": { | |
| "down_proj": { | |
| "absolute_error_sum": 5.391003262368031, | |
| "error_squared_sum": 0.008450353946089117, | |
| "max_absolute_error": 0.0058768130838871, | |
| "sample_values": 6144, | |
| "source_squared_sum": 0.9306811695457116 | |
| } | |
| }, | |
| "schema": "blackfrost.cyber_frost_nvfp4.shared_w13.v1", | |
| "shard_number": 86, | |
| "source": "model-00086-of-00131.safetensors", | |
| "storage_mode": "requantized_routed_experts_preserved_other_source_tensors" | |
| }, | |
| { | |
| "calibrated_gate_up_inputs_equal": true, | |
| "calibrated_input_scales_preserved_exactly": true, | |
| "elapsed_seconds": 10.659, | |
| "emitted_expert_matrices": 1024, | |
| "finite_positive_global_and_block_scales": true, | |
| "fused_source_tensors": 1, | |
| "gate_up_pairs_checked": 512, | |
| "output": "model-00087-of-00131.safetensors", | |
| "output_bytes": 944268824, | |
| "output_keys": 4096, | |
| "sample_reconstruction": { | |
| "gate_proj": { | |
| "absolute_error_sum": 5.5528873016592115, | |
| "error_squared_sum": 0.00884760515375391, | |
| "max_absolute_error": 0.0070626400411129, | |
| "sample_values": 6144, | |
| "source_squared_sum": 0.9665492573838632 | |
| }, | |
| "up_proj": { | |
| "absolute_error_sum": 5.676032300922088, | |
| "error_squared_sum": 0.009291482208762332, | |
| "max_absolute_error": 0.0064871646463871, | |
| "sample_values": 6144, | |
| "source_squared_sum": 1.0605005676893264 | |
| } | |
| }, | |
| "schema": "blackfrost.cyber_frost_nvfp4.shared_w13.v1", | |
| "shard_number": 87, | |
| "source": "model-00087-of-00131.safetensors", | |
| "storage_mode": "requantized_routed_experts_preserved_other_source_tensors" | |
| }, | |
| { | |
| "calibrated_gate_up_inputs_equal": true, | |
| "calibrated_input_scales_preserved_exactly": true, | |
| "elapsed_seconds": 4.325, | |
| "emitted_expert_matrices": 512, | |
| "finite_positive_global_and_block_scales": true, | |
| "fused_source_tensors": 1, | |
| "gate_up_pairs_checked": 0, | |
| "output": "model-00088-of-00131.safetensors", | |
| "output_bytes": 633487448, | |
| "output_keys": 2071, | |
| "sample_reconstruction": { | |
| "down_proj": { | |
| "absolute_error_sum": 5.597564881434664, | |
| "error_squared_sum": 0.009064361422523658, | |
| "max_absolute_error": 0.006103515625, | |
| "sample_values": 6144, | |
| "source_squared_sum": 0.9741301176729993 | |
| } | |
| }, | |
| "schema": "blackfrost.cyber_frost_nvfp4.shared_w13.v1", | |
| "shard_number": 88, | |
| "source": "model-00088-of-00131.safetensors", | |
| "storage_mode": "requantized_routed_experts_preserved_other_source_tensors" | |
| }, | |
| { | |
| "calibrated_gate_up_inputs_equal": true, | |
| "calibrated_input_scales_preserved_exactly": true, | |
| "elapsed_seconds": 10.936, | |
| "emitted_expert_matrices": 1024, | |
| "finite_positive_global_and_block_scales": true, | |
| "fused_source_tensors": 1, | |
| "gate_up_pairs_checked": 512, | |
| "output": "model-00089-of-00131.safetensors", | |
| "output_bytes": 944268824, | |
| "output_keys": 4096, | |
| "sample_reconstruction": { | |
| "gate_proj": { | |
| "absolute_error_sum": 5.522687681950629, | |
| "error_squared_sum": 0.008845328136911778, | |
| "max_absolute_error": 0.0061965212225914, | |
| "sample_values": 6144, | |
| "source_squared_sum": 0.9838946141423364 | |
| }, | |
| "up_proj": { | |
| "absolute_error_sum": 5.510751710273325, | |
| "error_squared_sum": 0.008838920226151275, | |
| "max_absolute_error": 0.007097512483596802, | |
| "sample_values": 6144, | |
| "source_squared_sum": 0.9762615595256765 | |
| } | |
| }, | |
| "schema": "blackfrost.cyber_frost_nvfp4.shared_w13.v1", | |
| "shard_number": 89, | |
| "source": "model-00089-of-00131.safetensors", | |
| "storage_mode": "requantized_routed_experts_preserved_other_source_tensors" | |
| }, | |
| { | |
| "calibrated_gate_up_inputs_equal": true, | |
| "calibrated_input_scales_preserved_exactly": true, | |
| "elapsed_seconds": 4.449, | |
| "emitted_expert_matrices": 512, | |
| "finite_positive_global_and_block_scales": true, | |
| "fused_source_tensors": 1, | |
| "gate_up_pairs_checked": 0, | |
| "output": "model-00090-of-00131.safetensors", | |
| "output_bytes": 630210640, | |
| "output_keys": 2071, | |
| "sample_reconstruction": { | |
| "down_proj": { | |
| "absolute_error_sum": 5.409989515203051, | |
| "error_squared_sum": 0.008633935627997559, | |
| "max_absolute_error": 0.007010322064161301, | |
| "sample_values": 6144, | |
| "source_squared_sum": 0.9699099960086219 | |
| } | |
| }, | |
| "schema": "blackfrost.cyber_frost_nvfp4.shared_w13.v1", | |
| "shard_number": 90, | |
| "source": "model-00090-of-00131.safetensors", | |
| "storage_mode": "requantized_routed_experts_preserved_other_source_tensors" | |
| }, | |
| { | |
| "calibrated_gate_up_inputs_equal": true, | |
| "calibrated_input_scales_preserved_exactly": true, | |
| "elapsed_seconds": 10.745, | |
| "emitted_expert_matrices": 1024, | |
| "finite_positive_global_and_block_scales": true, | |
| "fused_source_tensors": 1, | |
| "gate_up_pairs_checked": 512, | |
| "output": "model-00091-of-00131.safetensors", | |
| "output_bytes": 944268824, | |
| "output_keys": 4096, | |
| "sample_reconstruction": { | |
| "gate_proj": { | |
| "absolute_error_sum": 5.587539940373972, | |
| "error_squared_sum": 0.008914362878597617, | |
| "max_absolute_error": 0.0068359375, | |
| "sample_values": 6144, | |
| "source_squared_sum": 1.009994306117184 | |
| }, | |
| "up_proj": { | |
| "absolute_error_sum": 5.641103380476125, | |
| "error_squared_sum": 0.009232968607263636, | |
| "max_absolute_error": 0.0060744509100914, | |
| "sample_values": 6144, | |
| "source_squared_sum": 1.0263474993610453 | |
| } | |
| }, | |
| "schema": "blackfrost.cyber_frost_nvfp4.shared_w13.v1", | |
| "shard_number": 91, | |
| "source": "model-00091-of-00131.safetensors", | |
| "storage_mode": "requantized_routed_experts_preserved_other_source_tensors" | |
| }, | |
| { | |
| "calibrated_gate_up_inputs_equal": true, | |
| "calibrated_input_scales_preserved_exactly": true, | |
| "elapsed_seconds": 4.203, | |
| "emitted_expert_matrices": 512, | |
| "finite_positive_global_and_block_scales": true, | |
| "fused_source_tensors": 1, | |
| "gate_up_pairs_checked": 0, | |
| "output": "model-00092-of-00131.safetensors", | |
| "output_bytes": 511014256, | |
| "output_keys": 2061, | |
| "sample_reconstruction": { | |
| "down_proj": { | |
| "absolute_error_sum": 5.5660374944563955, | |
| "error_squared_sum": 0.008864927082566948, | |
| "max_absolute_error": 0.0056326724588871, | |
| "sample_values": 6144, | |
| "source_squared_sum": 1.0275229117560507 | |
| } | |
| }, | |
| "schema": "blackfrost.cyber_frost_nvfp4.shared_w13.v1", | |
| "shard_number": 92, | |
| "source": "model-00092-of-00131.safetensors", | |
| "storage_mode": "requantized_routed_experts_preserved_other_source_tensors" | |
| }, | |
| { | |
| "calibrated_gate_up_inputs_equal": true, | |
| "calibrated_input_scales_preserved_exactly": true, | |
| "elapsed_seconds": 10.82, | |
| "emitted_expert_matrices": 1024, | |
| "finite_positive_global_and_block_scales": true, | |
| "fused_source_tensors": 1, | |
| "gate_up_pairs_checked": 512, | |
| "output": "model-00093-of-00131.safetensors", | |
| "output_bytes": 944268824, | |
| "output_keys": 4096, | |
| "sample_reconstruction": { | |
| "gate_proj": { | |
| "absolute_error_sum": 6.016573155997321, | |
| "error_squared_sum": 0.010402702442648131, | |
| "max_absolute_error": 0.006940573453903198, | |
| "sample_values": 6144, | |
| "source_squared_sum": 1.1470707106905371 | |
| }, | |
| "up_proj": { | |
| "absolute_error_sum": 5.933673395891674, | |
| "error_squared_sum": 0.010305474153290008, | |
| "max_absolute_error": 0.006452292203903198, | |
| "sample_values": 6144, | |
| "source_squared_sum": 1.1649722880463624 | |
| } | |
| }, | |
| "schema": "blackfrost.cyber_frost_nvfp4.shared_w13.v1", | |
| "shard_number": 93, | |
| "source": "model-00093-of-00131.safetensors", | |
| "storage_mode": "requantized_routed_experts_preserved_other_source_tensors" | |
| }, | |
| { | |
| "calibrated_gate_up_inputs_equal": true, | |
| "calibrated_input_scales_preserved_exactly": true, | |
| "elapsed_seconds": 4.508, | |
| "emitted_expert_matrices": 512, | |
| "finite_positive_global_and_block_scales": true, | |
| "fused_source_tensors": 1, | |
| "gate_up_pairs_checked": 0, | |
| "output": "model-00094-of-00131.safetensors", | |
| "output_bytes": 729828344, | |
| "output_keys": 2080, | |
| "sample_reconstruction": { | |
| "down_proj": { | |
| "absolute_error_sum": 5.9609115773346275, | |
| "error_squared_sum": 0.01034361283712563, | |
| "max_absolute_error": 0.0076032355427742004, | |
| "sample_values": 6144, | |
| "source_squared_sum": 1.1423654677843382 | |
| } | |
| }, | |
| "schema": "blackfrost.cyber_frost_nvfp4.shared_w13.v1", | |
| "shard_number": 94, | |
| "source": "model-00094-of-00131.safetensors", | |
| "storage_mode": "requantized_routed_experts_preserved_other_source_tensors" | |
| }, | |
| { | |
| "calibrated_gate_up_inputs_equal": true, | |
| "calibrated_input_scales_preserved_exactly": true, | |
| "elapsed_seconds": 10.811, | |
| "emitted_expert_matrices": 1024, | |
| "finite_positive_global_and_block_scales": true, | |
| "fused_source_tensors": 1, | |
| "gate_up_pairs_checked": 512, | |
| "output": "model-00095-of-00131.safetensors", | |
| "output_bytes": 944268824, | |
| "output_keys": 4096, | |
| "sample_reconstruction": { | |
| "gate_proj": { | |
| "absolute_error_sum": 5.6660041401628405, | |
| "error_squared_sum": 0.009210422382404122, | |
| "max_absolute_error": 0.0064697265625, | |
| "sample_values": 6144, | |
| "source_squared_sum": 1.0430390696032548 | |
| }, | |
| "up_proj": { | |
| "absolute_error_sum": 5.623591054463759, | |
| "error_squared_sum": 0.009039353796317252, | |
| "max_absolute_error": 0.0059814453125, | |
| "sample_values": 6144, | |
| "source_squared_sum": 1.023509464202328 | |
| } | |
| }, | |
| "schema": "blackfrost.cyber_frost_nvfp4.shared_w13.v1", | |
| "shard_number": 95, | |
| "source": "model-00095-of-00131.safetensors", | |
| "storage_mode": "requantized_routed_experts_preserved_other_source_tensors" | |
| }, | |
| { | |
| "calibrated_gate_up_inputs_equal": true, | |
| "calibrated_input_scales_preserved_exactly": true, | |
| "elapsed_seconds": 4.543, | |
| "emitted_expert_matrices": 512, | |
| "finite_positive_global_and_block_scales": true, | |
| "fused_source_tensors": 1, | |
| "gate_up_pairs_checked": 0, | |
| "output": "model-00096-of-00131.safetensors", | |
| "output_bytes": 626933712, | |
| "output_keys": 2070, | |
| "sample_reconstruction": { | |
| "down_proj": { | |
| "absolute_error_sum": 5.652510441141203, | |
| "error_squared_sum": 0.009204063886772803, | |
| "max_absolute_error": 0.0069754458963871, | |
| "sample_values": 6144, | |
| "source_squared_sum": 1.0196119688189766 | |
| } | |
| }, | |
| "schema": "blackfrost.cyber_frost_nvfp4.shared_w13.v1", | |
| "shard_number": 96, | |
| "source": "model-00096-of-00131.safetensors", | |
| "storage_mode": "requantized_routed_experts_preserved_other_source_tensors" | |
| }, | |
| { | |
| "calibrated_gate_up_inputs_equal": true, | |
| "calibrated_input_scales_preserved_exactly": true, | |
| "elapsed_seconds": 11.213, | |
| "emitted_expert_matrices": 1024, | |
| "finite_positive_global_and_block_scales": true, | |
| "fused_source_tensors": 1, | |
| "gate_up_pairs_checked": 512, | |
| "output": "model-00097-of-00131.safetensors", | |
| "output_bytes": 944268824, | |
| "output_keys": 4096, | |
| "sample_reconstruction": { | |
| "gate_proj": { | |
| "absolute_error_sum": 5.79713650746271, | |
| "error_squared_sum": 0.009562765745914968, | |
| "max_absolute_error": 0.006591796875, | |
| "sample_values": 6144, | |
| "source_squared_sum": 1.0927413722279362 | |
| }, | |
| "up_proj": { | |
| "absolute_error_sum": 5.646686454303563, | |
| "error_squared_sum": 0.009157772034869397, | |
| "max_absolute_error": 0.00577799417078495, | |
| "sample_values": 6144, | |
| "source_squared_sum": 1.0054294435700644 | |
| } | |
| }, | |
| "schema": "blackfrost.cyber_frost_nvfp4.shared_w13.v1", | |
| "shard_number": 97, | |
| "source": "model-00097-of-00131.safetensors", | |
| "storage_mode": "requantized_routed_experts_preserved_other_source_tensors" | |
| }, | |
| { | |
| "calibrated_gate_up_inputs_equal": true, | |
| "calibrated_input_scales_preserved_exactly": true, | |
| "elapsed_seconds": 4.439, | |
| "emitted_expert_matrices": 512, | |
| "finite_positive_global_and_block_scales": true, | |
| "fused_source_tensors": 1, | |
| "gate_up_pairs_checked": 0, | |
| "output": "model-00098-of-00131.safetensors", | |
| "output_bytes": 626933712, | |
| "output_keys": 2070, | |
| "sample_reconstruction": { | |
| "down_proj": { | |
| "absolute_error_sum": 5.488481667125598, | |
| "error_squared_sum": 0.00866731952164584, | |
| "max_absolute_error": 0.005684986710548401, | |
| "sample_values": 6144, | |
| "source_squared_sum": 1.0008460367346024 | |
| } | |
| }, | |
| "schema": "blackfrost.cyber_frost_nvfp4.shared_w13.v1", | |
| "shard_number": 98, | |
| "source": "model-00098-of-00131.safetensors", | |
| "storage_mode": "requantized_routed_experts_preserved_other_source_tensors" | |
| }, | |
| { | |
| "calibrated_gate_up_inputs_equal": true, | |
| "calibrated_input_scales_preserved_exactly": true, | |
| "elapsed_seconds": 10.711, | |
| "emitted_expert_matrices": 1024, | |
| "finite_positive_global_and_block_scales": true, | |
| "fused_source_tensors": 1, | |
| "gate_up_pairs_checked": 512, | |
| "output": "model-00099-of-00131.safetensors", | |
| "output_bytes": 944268824, | |
| "output_keys": 4096, | |
| "sample_reconstruction": { | |
| "gate_proj": { | |
| "absolute_error_sum": 5.687156508560292, | |
| "error_squared_sum": 0.009557924732787215, | |
| "max_absolute_error": 0.0066964291036129, | |
| "sample_values": 6144, | |
| "source_squared_sum": 1.0421715144728019 | |
| }, | |
| "up_proj": { | |
| "absolute_error_sum": 5.578108275309205, | |
| "error_squared_sum": 0.009027798631049862, | |
| "max_absolute_error": 0.0062081478536129, | |
| "sample_values": 6144, | |
| "source_squared_sum": 1.0252065837079605 | |
| } | |
| }, | |
| "schema": "blackfrost.cyber_frost_nvfp4.shared_w13.v1", | |
| "shard_number": 99, | |
| "source": "model-00099-of-00131.safetensors", | |
| "storage_mode": "requantized_routed_experts_preserved_other_source_tensors" | |
| }, | |
| { | |
| "calibrated_gate_up_inputs_equal": true, | |
| "calibrated_input_scales_preserved_exactly": true, | |
| "elapsed_seconds": 4.25, | |
| "emitted_expert_matrices": 512, | |
| "finite_positive_global_and_block_scales": true, | |
| "fused_source_tensors": 1, | |
| "gate_up_pairs_checked": 0, | |
| "output": "model-00100-of-00131.safetensors", | |
| "output_bytes": 511014632, | |
| "output_keys": 2062, | |
| "sample_reconstruction": { | |
| "down_proj": { | |
| "absolute_error_sum": 5.6463828585110605, | |
| "error_squared_sum": 0.009262990710309296, | |
| "max_absolute_error": 0.0068243108689785, | |
| "sample_values": 6144, | |
| "source_squared_sum": 1.0389629922475925 | |
| } | |
| }, | |
| "schema": "blackfrost.cyber_frost_nvfp4.shared_w13.v1", | |
| "shard_number": 100, | |
| "source": "model-00100-of-00131.safetensors", | |
| "storage_mode": "requantized_routed_experts_preserved_other_source_tensors" | |
| }, | |
| { | |
| "calibrated_gate_up_inputs_equal": true, | |
| "calibrated_input_scales_preserved_exactly": true, | |
| "elapsed_seconds": 10.298, | |
| "emitted_expert_matrices": 1024, | |
| "finite_positive_global_and_block_scales": true, | |
| "fused_source_tensors": 1, | |
| "gate_up_pairs_checked": 512, | |
| "output": "model-00101-of-00131.safetensors", | |
| "output_bytes": 944268824, | |
| "output_keys": 4096, | |
| "sample_reconstruction": { | |
| "gate_proj": { | |
| "absolute_error_sum": 5.651539865764789, | |
| "error_squared_sum": 0.009304387148654409, | |
| "max_absolute_error": 0.0077078677713871, | |
| "sample_values": 6144, | |
| "source_squared_sum": 1.027388616012469 | |
| }, | |
| "up_proj": { | |
| "absolute_error_sum": 5.645827671280131, | |
| "error_squared_sum": 0.009292362405817645, | |
| "max_absolute_error": 0.0073707215487957, | |
| "sample_values": 6144, | |
| "source_squared_sum": 1.029903685856934 | |
| } | |
| }, | |
| "schema": "blackfrost.cyber_frost_nvfp4.shared_w13.v1", | |
| "shard_number": 101, | |
| "source": "model-00101-of-00131.safetensors", | |
| "storage_mode": "requantized_routed_experts_preserved_other_source_tensors" | |
| }, | |
| { | |
| "calibrated_gate_up_inputs_equal": true, | |
| "calibrated_input_scales_preserved_exactly": true, | |
| "elapsed_seconds": 4.549, | |
| "emitted_expert_matrices": 512, | |
| "finite_positive_global_and_block_scales": true, | |
| "fused_source_tensors": 1, | |
| "gate_up_pairs_checked": 0, | |
| "output": "model-00102-of-00131.safetensors", | |
| "output_bytes": 729828576, | |
| "output_keys": 2080, | |
| "sample_reconstruction": { | |
| "down_proj": { | |
| "absolute_error_sum": 5.651411758968607, | |
| "error_squared_sum": 0.009358829534860585, | |
| "max_absolute_error": 0.006417412310838699, | |
| "sample_values": 6144, | |
| "source_squared_sum": 1.0065969209882908 | |
| } | |
| }, | |
| "schema": "blackfrost.cyber_frost_nvfp4.shared_w13.v1", | |
| "shard_number": 102, | |
| "source": "model-00102-of-00131.safetensors", | |
| "storage_mode": "requantized_routed_experts_preserved_other_source_tensors" | |
| }, | |
| { | |
| "calibrated_gate_up_inputs_equal": true, | |
| "calibrated_input_scales_preserved_exactly": true, | |
| "elapsed_seconds": 11.671, | |
| "emitted_expert_matrices": 1024, | |
| "finite_positive_global_and_block_scales": true, | |
| "fused_source_tensors": 1, | |
| "gate_up_pairs_checked": 512, | |
| "output": "model-00103-of-00131.safetensors", | |
| "output_bytes": 944264728, | |
| "output_keys": 4096, | |
| "sample_reconstruction": { | |
| "gate_proj": { | |
| "absolute_error_sum": 5.618473636102863, | |
| "error_squared_sum": 0.009057471877950952, | |
| "max_absolute_error": 0.006295338273048401, | |
| "sample_values": 6144, | |
| "source_squared_sum": 1.028658020060598 | |
| }, | |
| "up_proj": { | |
| "absolute_error_sum": 5.579863375984132, | |
| "error_squared_sum": 0.009023958755973987, | |
| "max_absolute_error": 0.0063825324177742004, | |
| "sample_values": 6144, | |
| "source_squared_sum": 0.9816150444978698 | |
| } | |
| }, | |
| "schema": "blackfrost.cyber_frost_nvfp4.shared_w13.v1", | |
| "shard_number": 103, | |
| "source": "model-00103-of-00131.safetensors", | |
| "storage_mode": "requantized_routed_experts_preserved_other_source_tensors" | |
| }, | |
| { | |
| "calibrated_gate_up_inputs_equal": true, | |
| "calibrated_input_scales_preserved_exactly": true, | |
| "elapsed_seconds": 4.66, | |
| "emitted_expert_matrices": 512, | |
| "finite_positive_global_and_block_scales": true, | |
| "fused_source_tensors": 1, | |
| "gate_up_pairs_checked": 0, | |
| "output": "model-00104-of-00131.safetensors", | |
| "output_bytes": 626931656, | |
| "output_keys": 2070, | |
| "sample_reconstruction": { | |
| "down_proj": { | |
| "absolute_error_sum": 5.4863899749470875, | |
| "error_squared_sum": 0.008767387234907657, | |
| "max_absolute_error": 0.0064522884786129, | |
| "sample_values": 6144, | |
| "source_squared_sum": 0.968660518686149 | |
| } | |
| }, | |
| "schema": "blackfrost.cyber_frost_nvfp4.shared_w13.v1", | |
| "shard_number": 104, | |
| "source": "model-00104-of-00131.safetensors", | |
| "storage_mode": "requantized_routed_experts_preserved_other_source_tensors" | |
| }, | |
| { | |
| "calibrated_gate_up_inputs_equal": true, | |
| "calibrated_input_scales_preserved_exactly": true, | |
| "elapsed_seconds": 10.786, | |
| "emitted_expert_matrices": 1024, | |
| "finite_positive_global_and_block_scales": true, | |
| "fused_source_tensors": 1, | |
| "gate_up_pairs_checked": 512, | |
| "output": "model-00105-of-00131.safetensors", | |
| "output_bytes": 944268824, | |
| "output_keys": 4096, | |
| "sample_reconstruction": { | |
| "gate_proj": { | |
| "absolute_error_sum": 6.049879088997841, | |
| "error_squared_sum": 0.010775793546398165, | |
| "max_absolute_error": 0.007254466414451599, | |
| "sample_values": 6144, | |
| "source_squared_sum": 1.1582036673291443 | |
| }, | |
| "up_proj": { | |
| "absolute_error_sum": 5.8380125312833115, | |
| "error_squared_sum": 0.00982096540041058, | |
| "max_absolute_error": 0.00582449696958065, | |
| "sample_values": 6144, | |
| "source_squared_sum": 1.0706816263756154 | |
| } | |
| }, | |
| "schema": "blackfrost.cyber_frost_nvfp4.shared_w13.v1", | |
| "shard_number": 105, | |
| "source": "model-00105-of-00131.safetensors", | |
| "storage_mode": "requantized_routed_experts_preserved_other_source_tensors" | |
| }, | |
| { | |
| "calibrated_gate_up_inputs_equal": true, | |
| "calibrated_input_scales_preserved_exactly": true, | |
| "elapsed_seconds": 4.33, | |
| "emitted_expert_matrices": 512, | |
| "finite_positive_global_and_block_scales": true, | |
| "fused_source_tensors": 1, | |
| "gate_up_pairs_checked": 0, | |
| "output": "model-00106-of-00131.safetensors", | |
| "output_bytes": 658391104, | |
| "output_keys": 2071, | |
| "sample_reconstruction": { | |
| "down_proj": { | |
| "absolute_error_sum": 5.47671317338245, | |
| "error_squared_sum": 0.008749487110711045, | |
| "max_absolute_error": 0.005929127335548401, | |
| "sample_values": 6144, | |
| "source_squared_sum": 0.976272796830262 | |
| } | |
| }, | |
| "schema": "blackfrost.cyber_frost_nvfp4.shared_w13.v1", | |
| "shard_number": 106, | |
| "source": "model-00106-of-00131.safetensors", | |
| "storage_mode": "requantized_routed_experts_preserved_other_source_tensors" | |
| }, | |
| { | |
| "calibrated_gate_up_inputs_equal": true, | |
| "calibrated_input_scales_preserved_exactly": true, | |
| "elapsed_seconds": 11.007, | |
| "emitted_expert_matrices": 1024, | |
| "finite_positive_global_and_block_scales": true, | |
| "fused_source_tensors": 1, | |
| "gate_up_pairs_checked": 512, | |
| "output": "model-00107-of-00131.safetensors", | |
| "output_bytes": 944268824, | |
| "output_keys": 4096, | |
| "sample_reconstruction": { | |
| "gate_proj": { | |
| "absolute_error_sum": 5.919292525504716, | |
| "error_squared_sum": 0.010141408603915152, | |
| "max_absolute_error": 0.006766181439161301, | |
| "sample_values": 6144, | |
| "source_squared_sum": 1.1388920026970217 | |
| }, | |
| "up_proj": { | |
| "absolute_error_sum": 5.638837523176335, | |
| "error_squared_sum": 0.009198321632450365, | |
| "max_absolute_error": 0.00647553987801075, | |
| "sample_values": 6144, | |
| "source_squared_sum": 1.0444109003843938 | |
| } | |
| }, | |
| "schema": "blackfrost.cyber_frost_nvfp4.shared_w13.v1", | |
| "shard_number": 107, | |
| "source": "model-00107-of-00131.safetensors", | |
| "storage_mode": "requantized_routed_experts_preserved_other_source_tensors" | |
| }, | |
| { | |
| "calibrated_gate_up_inputs_equal": true, | |
| "calibrated_input_scales_preserved_exactly": true, | |
| "elapsed_seconds": 4.505, | |
| "emitted_expert_matrices": 512, | |
| "finite_positive_global_and_block_scales": true, | |
| "fused_source_tensors": 1, | |
| "gate_up_pairs_checked": 0, | |
| "output": "model-00108-of-00131.safetensors", | |
| "output_bytes": 695091488, | |
| "output_keys": 2073, | |
| "sample_reconstruction": { | |
| "down_proj": { | |
| "absolute_error_sum": 5.396944243344478, | |
| "error_squared_sum": 0.008480858821372193, | |
| "max_absolute_error": 0.007498607039451599, | |
| "sample_values": 6144, | |
| "source_squared_sum": 0.9603729188260532 | |
| } | |
| }, | |
| "schema": "blackfrost.cyber_frost_nvfp4.shared_w13.v1", | |
| "shard_number": 108, | |
| "source": "model-00108-of-00131.safetensors", | |
| "storage_mode": "requantized_routed_experts_preserved_other_source_tensors" | |
| }, | |
| { | |
| "calibrated_gate_up_inputs_equal": true, | |
| "calibrated_input_scales_preserved_exactly": true, | |
| "elapsed_seconds": 10.592, | |
| "emitted_expert_matrices": 1024, | |
| "finite_positive_global_and_block_scales": true, | |
| "fused_source_tensors": 1, | |
| "gate_up_pairs_checked": 512, | |
| "output": "model-00109-of-00131.safetensors", | |
| "output_bytes": 944268824, | |
| "output_keys": 4096, | |
| "sample_reconstruction": { | |
| "gate_proj": { | |
| "absolute_error_sum": 5.9547989261336625, | |
| "error_squared_sum": 0.010187559734898276, | |
| "max_absolute_error": 0.0066266730427742004, | |
| "sample_values": 6144, | |
| "source_squared_sum": 1.136262825529953 | |
| }, | |
| "up_proj": { | |
| "absolute_error_sum": 5.841407991480082, | |
| "error_squared_sum": 0.009751280555290563, | |
| "max_absolute_error": 0.005998879671096802, | |
| "sample_values": 6144, | |
| "source_squared_sum": 1.0942883796258167 | |
| } | |
| }, | |
| "schema": "blackfrost.cyber_frost_nvfp4.shared_w13.v1", | |
| "shard_number": 109, | |
| "source": "model-00109-of-00131.safetensors", | |
| "storage_mode": "requantized_routed_experts_preserved_other_source_tensors" | |
| }, | |
| { | |
| "calibrated_gate_up_inputs_equal": true, | |
| "calibrated_input_scales_preserved_exactly": true, | |
| "elapsed_seconds": 4.229, | |
| "emitted_expert_matrices": 512, | |
| "finite_positive_global_and_block_scales": true, | |
| "fused_source_tensors": 1, | |
| "gate_up_pairs_checked": 0, | |
| "output": "model-00110-of-00131.safetensors", | |
| "output_bytes": 511014256, | |
| "output_keys": 2061, | |
| "sample_reconstruction": { | |
| "down_proj": { | |
| "absolute_error_sum": 5.754020327585749, | |
| "error_squared_sum": 0.009466242170039567, | |
| "max_absolute_error": 0.005615234375, | |
| "sample_values": 6144, | |
| "source_squared_sum": 1.0640974317872582 | |
| } | |
| }, | |
| "schema": "blackfrost.cyber_frost_nvfp4.shared_w13.v1", | |
| "shard_number": 110, | |
| "source": "model-00110-of-00131.safetensors", | |
| "storage_mode": "requantized_routed_experts_preserved_other_source_tensors" | |
| }, | |
| { | |
| "calibrated_gate_up_inputs_equal": true, | |
| "calibrated_input_scales_preserved_exactly": true, | |
| "elapsed_seconds": 10.81, | |
| "emitted_expert_matrices": 1024, | |
| "finite_positive_global_and_block_scales": true, | |
| "fused_source_tensors": 1, | |
| "gate_up_pairs_checked": 512, | |
| "output": "model-00111-of-00131.safetensors", | |
| "output_bytes": 944268824, | |
| "output_keys": 4096, | |
| "sample_reconstruction": { | |
| "gate_proj": { | |
| "absolute_error_sum": 5.608159610885195, | |
| "error_squared_sum": 0.009088323246757981, | |
| "max_absolute_error": 0.0061965212225914, | |
| "sample_values": 6144, | |
| "source_squared_sum": 0.996194992199805 | |
| }, | |
| "up_proj": { | |
| "absolute_error_sum": 5.5106119707925245, | |
| "error_squared_sum": 0.008583987108966974, | |
| "max_absolute_error": 0.00634765625, | |
| "sample_values": 6144, | |
| "source_squared_sum": 0.9803876172205046 | |
| } | |
| }, | |
| "schema": "blackfrost.cyber_frost_nvfp4.shared_w13.v1", | |
| "shard_number": 111, | |
| "source": "model-00111-of-00131.safetensors", | |
| "storage_mode": "requantized_routed_experts_preserved_other_source_tensors" | |
| }, | |
| { | |
| "calibrated_gate_up_inputs_equal": true, | |
| "calibrated_input_scales_preserved_exactly": true, | |
| "elapsed_seconds": 4.929, | |
| "emitted_expert_matrices": 512, | |
| "finite_positive_global_and_block_scales": true, | |
| "fused_source_tensors": 1, | |
| "gate_up_pairs_checked": 0, | |
| "output": "model-00112-of-00131.safetensors", | |
| "output_bytes": 729828584, | |
| "output_keys": 2080, | |
| "sample_reconstruction": { | |
| "down_proj": { | |
| "absolute_error_sum": 5.5866703724022955, | |
| "error_squared_sum": 0.009077224402430983, | |
| "max_absolute_error": 0.00690569169819355, | |
| "sample_values": 6144, | |
| "source_squared_sum": 1.034490622272827 | |
| } | |
| }, | |
| "schema": "blackfrost.cyber_frost_nvfp4.shared_w13.v1", | |
| "shard_number": 112, | |
| "source": "model-00112-of-00131.safetensors", | |
| "storage_mode": "requantized_routed_experts_preserved_other_source_tensors" | |
| }, | |
| { | |
| "calibrated_gate_up_inputs_equal": true, | |
| "calibrated_input_scales_preserved_exactly": true, | |
| "elapsed_seconds": 11.442, | |
| "emitted_expert_matrices": 1024, | |
| "finite_positive_global_and_block_scales": true, | |
| "fused_source_tensors": 1, | |
| "gate_up_pairs_checked": 512, | |
| "output": "model-00113-of-00131.safetensors", | |
| "output_bytes": 944268824, | |
| "output_keys": 4096, | |
| "sample_reconstruction": { | |
| "gate_proj": { | |
| "absolute_error_sum": 5.823439210420474, | |
| "error_squared_sum": 0.009836023890486238, | |
| "max_absolute_error": 0.007080078125, | |
| "sample_values": 6144, | |
| "source_squared_sum": 1.093593330332709 | |
| }, | |
| "up_proj": { | |
| "absolute_error_sum": 5.770708601572551, | |
| "error_squared_sum": 0.009457138908231136, | |
| "max_absolute_error": 0.00626627542078495, | |
| "sample_values": 6144, | |
| "source_squared_sum": 1.0783553278179587 | |
| } | |
| }, | |
| "schema": "blackfrost.cyber_frost_nvfp4.shared_w13.v1", | |
| "shard_number": 113, | |
| "source": "model-00113-of-00131.safetensors", | |
| "storage_mode": "requantized_routed_experts_preserved_other_source_tensors" | |
| }, | |
| { | |
| "calibrated_gate_up_inputs_equal": true, | |
| "calibrated_input_scales_preserved_exactly": true, | |
| "elapsed_seconds": 4.365, | |
| "emitted_expert_matrices": 512, | |
| "finite_positive_global_and_block_scales": true, | |
| "fused_source_tensors": 1, | |
| "gate_up_pairs_checked": 0, | |
| "output": "model-00114-of-00131.safetensors", | |
| "output_bytes": 626938936, | |
| "output_keys": 2071, | |
| "sample_reconstruction": { | |
| "down_proj": { | |
| "absolute_error_sum": 5.857965215225704, | |
| "error_squared_sum": 0.00986836120046317, | |
| "max_absolute_error": 0.0068010613322257996, | |
| "sample_values": 6144, | |
| "source_squared_sum": 1.1081167137579244 | |
| } | |
| }, | |
| "schema": "blackfrost.cyber_frost_nvfp4.shared_w13.v1", | |
| "shard_number": 114, | |
| "source": "model-00114-of-00131.safetensors", | |
| "storage_mode": "requantized_routed_experts_preserved_other_source_tensors" | |
| }, | |
| { | |
| "calibrated_gate_up_inputs_equal": true, | |
| "calibrated_input_scales_preserved_exactly": true, | |
| "elapsed_seconds": 10.401, | |
| "emitted_expert_matrices": 1024, | |
| "finite_positive_global_and_block_scales": true, | |
| "fused_source_tensors": 1, | |
| "gate_up_pairs_checked": 512, | |
| "output": "model-00115-of-00131.safetensors", | |
| "output_bytes": 944268824, | |
| "output_keys": 4096, | |
| "sample_reconstruction": { | |
| "gate_proj": { | |
| "absolute_error_sum": 5.738162201829255, | |
| "error_squared_sum": 0.009574350275859287, | |
| "max_absolute_error": 0.00592912919819355, | |
| "sample_values": 6144, | |
| "source_squared_sum": 1.0504893849626393 | |
| }, | |
| "up_proj": { | |
| "absolute_error_sum": 5.462528009200469, | |
| "error_squared_sum": 0.008617321548914324, | |
| "max_absolute_error": 0.005929131060838699, | |
| "sample_values": 6144, | |
| "source_squared_sum": 0.9758777648394203 | |
| } | |
| }, | |
| "schema": "blackfrost.cyber_frost_nvfp4.shared_w13.v1", | |
| "shard_number": 115, | |
| "source": "model-00115-of-00131.safetensors", | |
| "storage_mode": "requantized_routed_experts_preserved_other_source_tensors" | |
| }, | |
| { | |
| "calibrated_gate_up_inputs_equal": true, | |
| "calibrated_input_scales_preserved_exactly": true, | |
| "elapsed_seconds": 4.266, | |
| "emitted_expert_matrices": 512, | |
| "finite_positive_global_and_block_scales": true, | |
| "fused_source_tensors": 1, | |
| "gate_up_pairs_checked": 0, | |
| "output": "model-00116-of-00131.safetensors", | |
| "output_bytes": 626933712, | |
| "output_keys": 2070, | |
| "sample_reconstruction": { | |
| "down_proj": { | |
| "absolute_error_sum": 5.6223027523374185, | |
| "error_squared_sum": 0.009024586773283477, | |
| "max_absolute_error": 0.0061965212225914, | |
| "sample_values": 6144, | |
| "source_squared_sum": 1.0294488413794283 | |
| } | |
| }, | |
| "schema": "blackfrost.cyber_frost_nvfp4.shared_w13.v1", | |
| "shard_number": 116, | |
| "source": "model-00116-of-00131.safetensors", | |
| "storage_mode": "requantized_routed_experts_preserved_other_source_tensors" | |
| }, | |
| { | |
| "calibrated_gate_up_inputs_equal": true, | |
| "calibrated_input_scales_preserved_exactly": true, | |
| "elapsed_seconds": 10.486, | |
| "emitted_expert_matrices": 1024, | |
| "finite_positive_global_and_block_scales": true, | |
| "fused_source_tensors": 1, | |
| "gate_up_pairs_checked": 512, | |
| "output": "model-00117-of-00131.safetensors", | |
| "output_bytes": 944268824, | |
| "output_keys": 4096, | |
| "sample_reconstruction": { | |
| "gate_proj": { | |
| "absolute_error_sum": 5.713086030445993, | |
| "error_squared_sum": 0.00942524931720812, | |
| "max_absolute_error": 0.0058942511677742004, | |
| "sample_values": 6144, | |
| "source_squared_sum": 1.0581106043315336 | |
| }, | |
| "up_proj": { | |
| "absolute_error_sum": 5.8065972769400105, | |
| "error_squared_sum": 0.009815816475793486, | |
| "max_absolute_error": 0.007498603314161301, | |
| "sample_values": 6144, | |
| "source_squared_sum": 1.1032407132633844 | |
| } | |
| }, | |
| "schema": "blackfrost.cyber_frost_nvfp4.shared_w13.v1", | |
| "shard_number": 117, | |
| "source": "model-00117-of-00131.safetensors", | |
| "storage_mode": "requantized_routed_experts_preserved_other_source_tensors" | |
| }, | |
| { | |
| "calibrated_gate_up_inputs_equal": true, | |
| "calibrated_input_scales_preserved_exactly": true, | |
| "elapsed_seconds": 4.148, | |
| "emitted_expert_matrices": 512, | |
| "finite_positive_global_and_block_scales": true, | |
| "fused_source_tensors": 1, | |
| "gate_up_pairs_checked": 0, | |
| "output": "model-00118-of-00131.safetensors", | |
| "output_bytes": 511034840, | |
| "output_keys": 2062, | |
| "sample_reconstruction": { | |
| "down_proj": { | |
| "absolute_error_sum": 5.863365433644503, | |
| "error_squared_sum": 0.009964706832097627, | |
| "max_absolute_error": 0.006766181439161301, | |
| "sample_values": 6144, | |
| "source_squared_sum": 1.109671887031822 | |
| } | |
| }, | |
| "schema": "blackfrost.cyber_frost_nvfp4.shared_w13.v1", | |
| "shard_number": 118, | |
| "source": "model-00118-of-00131.safetensors", | |
| "storage_mode": "requantized_routed_experts_preserved_other_source_tensors" | |
| }, | |
| { | |
| "calibrated_gate_up_inputs_equal": true, | |
| "calibrated_input_scales_preserved_exactly": true, | |
| "elapsed_seconds": 10.382, | |
| "emitted_expert_matrices": 1024, | |
| "finite_positive_global_and_block_scales": true, | |
| "fused_source_tensors": 1, | |
| "gate_up_pairs_checked": 512, | |
| "output": "model-00119-of-00131.safetensors", | |
| "output_bytes": 944268824, | |
| "output_keys": 4096, | |
| "sample_reconstruction": { | |
| "gate_proj": { | |
| "absolute_error_sum": 6.48432196083013, | |
| "error_squared_sum": 0.012545542466297408, | |
| "max_absolute_error": 0.008544921875, | |
| "sample_values": 6144, | |
| "source_squared_sum": 1.4003755790336774 | |
| }, | |
| "up_proj": { | |
| "absolute_error_sum": 6.533774282899685, | |
| "error_squared_sum": 0.012273439789753161, | |
| "max_absolute_error": 0.0078125, | |
| "sample_values": 6144, | |
| "source_squared_sum": 1.3887645319044684 | |
| } | |
| }, | |
| "schema": "blackfrost.cyber_frost_nvfp4.shared_w13.v1", | |
| "shard_number": 119, | |
| "source": "model-00119-of-00131.safetensors", | |
| "storage_mode": "requantized_routed_experts_preserved_other_source_tensors" | |
| }, | |
| { | |
| "calibrated_gate_up_inputs_equal": true, | |
| "calibrated_input_scales_preserved_exactly": true, | |
| "elapsed_seconds": 4.466, | |
| "emitted_expert_matrices": 512, | |
| "finite_positive_global_and_block_scales": true, | |
| "fused_source_tensors": 1, | |
| "gate_up_pairs_checked": 0, | |
| "output": "model-00120-of-00131.safetensors", | |
| "output_bytes": 742935256, | |
| "output_keys": 2080, | |
| "sample_reconstruction": { | |
| "down_proj": { | |
| "absolute_error_sum": 6.307892757235095, | |
| "error_squared_sum": 0.011577116477858609, | |
| "max_absolute_error": 0.0075334832072257996, | |
| "sample_values": 6144, | |
| "source_squared_sum": 1.2818349110480955 | |
| } | |
| }, | |
| "schema": "blackfrost.cyber_frost_nvfp4.shared_w13.v1", | |
| "shard_number": 120, | |
| "source": "model-00120-of-00131.safetensors", | |
| "storage_mode": "requantized_routed_experts_preserved_other_source_tensors" | |
| }, | |
| { | |
| "calibrated_gate_up_inputs_equal": true, | |
| "calibrated_input_scales_preserved_exactly": true, | |
| "elapsed_seconds": 10.894, | |
| "emitted_expert_matrices": 1024, | |
| "finite_positive_global_and_block_scales": true, | |
| "fused_source_tensors": 1, | |
| "gate_up_pairs_checked": 512, | |
| "output": "model-00121-of-00131.safetensors", | |
| "output_bytes": 944264728, | |
| "output_keys": 4096, | |
| "sample_reconstruction": { | |
| "gate_proj": { | |
| "absolute_error_sum": 5.803681372082792, | |
| "error_squared_sum": 0.009743076124610553, | |
| "max_absolute_error": 0.006277903914451599, | |
| "sample_values": 6144, | |
| "source_squared_sum": 1.057970941547815 | |
| }, | |
| "up_proj": { | |
| "absolute_error_sum": 5.78947907849215, | |
| "error_squared_sum": 0.009652409857391768, | |
| "max_absolute_error": 0.00632440485060215, | |
| "sample_values": 6144, | |
| "source_squared_sum": 1.0668761360944052 | |
| } | |
| }, | |
| "schema": "blackfrost.cyber_frost_nvfp4.shared_w13.v1", | |
| "shard_number": 121, | |
| "source": "model-00121-of-00131.safetensors", | |
| "storage_mode": "requantized_routed_experts_preserved_other_source_tensors" | |
| }, | |
| { | |
| "calibrated_gate_up_inputs_equal": true, | |
| "calibrated_input_scales_preserved_exactly": true, | |
| "elapsed_seconds": 4.562, | |
| "emitted_expert_matrices": 512, | |
| "finite_positive_global_and_block_scales": true, | |
| "fused_source_tensors": 1, | |
| "gate_up_pairs_checked": 0, | |
| "output": "model-00122-of-00131.safetensors", | |
| "output_bytes": 626931640, | |
| "output_keys": 2070, | |
| "sample_reconstruction": { | |
| "down_proj": { | |
| "absolute_error_sum": 5.318389473832212, | |
| "error_squared_sum": 0.00809274754882786, | |
| "max_absolute_error": 0.0059640072286129, | |
| "sample_values": 6144, | |
| "source_squared_sum": 0.9193412011589347 | |
| } | |
| }, | |
| "schema": "blackfrost.cyber_frost_nvfp4.shared_w13.v1", | |
| "shard_number": 122, | |
| "source": "model-00122-of-00131.safetensors", | |
| "storage_mode": "requantized_routed_experts_preserved_other_source_tensors" | |
| }, | |
| { | |
| "calibrated_gate_up_inputs_equal": true, | |
| "calibrated_input_scales_preserved_exactly": true, | |
| "elapsed_seconds": 10.57, | |
| "emitted_expert_matrices": 1024, | |
| "finite_positive_global_and_block_scales": true, | |
| "fused_source_tensors": 1, | |
| "gate_up_pairs_checked": 512, | |
| "output": "model-00123-of-00131.safetensors", | |
| "output_bytes": 944264728, | |
| "output_keys": 4096, | |
| "sample_reconstruction": { | |
| "gate_proj": { | |
| "absolute_error_sum": 5.589251919533126, | |
| "error_squared_sum": 0.008881100114157329, | |
| "max_absolute_error": 0.0060918889939785, | |
| "sample_values": 6144, | |
| "source_squared_sum": 0.9886275433588345 | |
| }, | |
| "up_proj": { | |
| "absolute_error_sum": 5.475439126137644, | |
| "error_squared_sum": 0.008699432667263956, | |
| "max_absolute_error": 0.0054582878947257996, | |
| "sample_values": 6144, | |
| "source_squared_sum": 0.9678340193924955 | |
| } | |
| }, | |
| "schema": "blackfrost.cyber_frost_nvfp4.shared_w13.v1", | |
| "shard_number": 123, | |
| "source": "model-00123-of-00131.safetensors", | |
| "storage_mode": "requantized_routed_experts_preserved_other_source_tensors" | |
| }, | |
| { | |
| "calibrated_gate_up_inputs_equal": true, | |
| "calibrated_input_scales_preserved_exactly": true, | |
| "elapsed_seconds": 4.038, | |
| "emitted_expert_matrices": 512, | |
| "finite_positive_global_and_block_scales": true, | |
| "fused_source_tensors": 1, | |
| "gate_up_pairs_checked": 0, | |
| "output": "model-00124-of-00131.safetensors", | |
| "output_bytes": 524119600, | |
| "output_keys": 2062, | |
| "sample_reconstruction": { | |
| "down_proj": { | |
| "absolute_error_sum": 5.295317433774471, | |
| "error_squared_sum": 0.008109238175290561, | |
| "max_absolute_error": 0.0055571049451828, | |
| "sample_values": 6144, | |
| "source_squared_sum": 0.9064775172067145 | |
| } | |
| }, | |
| "schema": "blackfrost.cyber_frost_nvfp4.shared_w13.v1", | |
| "shard_number": 124, | |
| "source": "model-00124-of-00131.safetensors", | |
| "storage_mode": "requantized_routed_experts_preserved_other_source_tensors" | |
| }, | |
| { | |
| "calibrated_gate_up_inputs_equal": true, | |
| "calibrated_input_scales_preserved_exactly": true, | |
| "elapsed_seconds": 10.47, | |
| "emitted_expert_matrices": 1024, | |
| "finite_positive_global_and_block_scales": true, | |
| "fused_source_tensors": 1, | |
| "gate_up_pairs_checked": 512, | |
| "output": "model-00125-of-00131.safetensors", | |
| "output_bytes": 944264728, | |
| "output_keys": 4096, | |
| "sample_reconstruction": { | |
| "gate_proj": { | |
| "absolute_error_sum": 5.619401631876826, | |
| "error_squared_sum": 0.0090340287801859, | |
| "max_absolute_error": 0.00690569169819355, | |
| "sample_values": 6144, | |
| "source_squared_sum": 1.0356122441727325 | |
| }, | |
| "up_proj": { | |
| "absolute_error_sum": 5.51413649902679, | |
| "error_squared_sum": 0.008882386764907827, | |
| "max_absolute_error": 0.00680105946958065, | |
| "sample_values": 6144, | |
| "source_squared_sum": 0.9764717890367804 | |
| } | |
| }, | |
| "schema": "blackfrost.cyber_frost_nvfp4.shared_w13.v1", | |
| "shard_number": 125, | |
| "source": "model-00125-of-00131.safetensors", | |
| "storage_mode": "requantized_routed_experts_preserved_other_source_tensors" | |
| }, | |
| { | |
| "calibrated_gate_up_inputs_equal": true, | |
| "calibrated_input_scales_preserved_exactly": true, | |
| "elapsed_seconds": 4.375, | |
| "emitted_expert_matrices": 512, | |
| "finite_positive_global_and_block_scales": true, | |
| "fused_source_tensors": 1, | |
| "gate_up_pairs_checked": 0, | |
| "output": "model-00126-of-00131.safetensors", | |
| "output_bytes": 729825888, | |
| "output_keys": 2079, | |
| "sample_reconstruction": { | |
| "down_proj": { | |
| "absolute_error_sum": 5.114019644388463, | |
| "error_squared_sum": 0.007578864207692569, | |
| "max_absolute_error": 0.007056824862957001, | |
| "sample_values": 6144, | |
| "source_squared_sum": 0.8408277999412335 | |
| } | |
| }, | |
| "schema": "blackfrost.cyber_frost_nvfp4.shared_w13.v1", | |
| "shard_number": 126, | |
| "source": "model-00126-of-00131.safetensors", | |
| "storage_mode": "requantized_routed_experts_preserved_other_source_tensors" | |
| }, | |
| { | |
| "calibrated_gate_up_inputs_equal": true, | |
| "calibrated_input_scales_preserved_exactly": true, | |
| "elapsed_seconds": 10.46, | |
| "emitted_expert_matrices": 1024, | |
| "finite_positive_global_and_block_scales": true, | |
| "fused_source_tensors": 1, | |
| "gate_up_pairs_checked": 512, | |
| "output": "model-00127-of-00131.safetensors", | |
| "output_bytes": 944264728, | |
| "output_keys": 4096, | |
| "sample_reconstruction": { | |
| "gate_proj": { | |
| "absolute_error_sum": 5.6527589990291744, | |
| "error_squared_sum": 0.009343047165560567, | |
| "max_absolute_error": 0.0060163214802742004, | |
| "sample_values": 6144, | |
| "source_squared_sum": 1.0005920123423486 | |
| }, | |
| "up_proj": { | |
| "absolute_error_sum": 5.544696036027744, | |
| "error_squared_sum": 0.008788178256701782, | |
| "max_absolute_error": 0.0066964291036129, | |
| "sample_values": 6144, | |
| "source_squared_sum": 1.010241752914084 | |
| } | |
| }, | |
| "schema": "blackfrost.cyber_frost_nvfp4.shared_w13.v1", | |
| "shard_number": 127, | |
| "source": "model-00127-of-00131.safetensors", | |
| "storage_mode": "requantized_routed_experts_preserved_other_source_tensors" | |
| }, | |
| { | |
| "calibrated_gate_up_inputs_equal": true, | |
| "calibrated_input_scales_preserved_exactly": true, | |
| "elapsed_seconds": 4.254, | |
| "emitted_expert_matrices": 512, | |
| "finite_positive_global_and_block_scales": true, | |
| "fused_source_tensors": 1, | |
| "gate_up_pairs_checked": 0, | |
| "output": "model-00128-of-00131.safetensors", | |
| "output_bytes": 626931640, | |
| "output_keys": 2070, | |
| "sample_reconstruction": { | |
| "down_proj": { | |
| "absolute_error_sum": 5.650970106595196, | |
| "error_squared_sum": 0.00926202030717399, | |
| "max_absolute_error": 0.006068635731935501, | |
| "sample_values": 6144, | |
| "source_squared_sum": 0.9956778073793293 | |
| } | |
| }, | |
| "schema": "blackfrost.cyber_frost_nvfp4.shared_w13.v1", | |
| "shard_number": 128, | |
| "source": "model-00128-of-00131.safetensors", | |
| "storage_mode": "requantized_routed_experts_preserved_other_source_tensors" | |
| }, | |
| { | |
| "calibrated_gate_up_inputs_equal": true, | |
| "calibrated_input_scales_preserved_exactly": true, | |
| "elapsed_seconds": 13.904, | |
| "emitted_expert_matrices": 1024, | |
| "finite_positive_global_and_block_scales": true, | |
| "fused_source_tensors": 1, | |
| "gate_up_pairs_checked": 512, | |
| "output": "model-00129-of-00131.safetensors", | |
| "output_bytes": 944264728, | |
| "output_keys": 4096, | |
| "sample_reconstruction": { | |
| "gate_proj": { | |
| "absolute_error_sum": 5.6441783520858735, | |
| "error_squared_sum": 0.009170366829112958, | |
| "max_absolute_error": 0.0059465691447257996, | |
| "sample_values": 6144, | |
| "source_squared_sum": 1.0028108816484194 | |
| }, | |
| "up_proj": { | |
| "absolute_error_sum": 5.440823545446619, | |
| "error_squared_sum": 0.008583523931371219, | |
| "max_absolute_error": 0.005615234375, | |
| "sample_values": 6144, | |
| "source_squared_sum": 0.9462550660899249 | |
| } | |
| }, | |
| "schema": "blackfrost.cyber_frost_nvfp4.shared_w13.v1", | |
| "shard_number": 129, | |
| "source": "model-00129-of-00131.safetensors", | |
| "storage_mode": "requantized_routed_experts_preserved_other_source_tensors" | |
| }, | |
| { | |
| "calibrated_gate_up_inputs_equal": true, | |
| "calibrated_input_scales_preserved_exactly": true, | |
| "elapsed_seconds": 6.107, | |
| "emitted_expert_matrices": 512, | |
| "finite_positive_global_and_block_scales": true, | |
| "fused_source_tensors": 1, | |
| "gate_up_pairs_checked": 0, | |
| "output": "model-00130-of-00131.safetensors", | |
| "output_bytes": 1769203656, | |
| "output_keys": 2058, | |
| "sample_reconstruction": { | |
| "down_proj": { | |
| "absolute_error_sum": 5.672889996669255, | |
| "error_squared_sum": 0.009357194490799671, | |
| "max_absolute_error": 0.0071149542927742004, | |
| "sample_values": 6144, | |
| "source_squared_sum": 1.038680192212114 | |
| } | |
| }, | |
| "schema": "blackfrost.cyber_frost_nvfp4.shared_w13.v1", | |
| "shard_number": 130, | |
| "source": "model-00130-of-00131.safetensors", | |
| "storage_mode": "requantized_routed_experts_preserved_other_source_tensors" | |
| }, | |
| { | |
| "calibrated_gate_up_inputs_equal": true, | |
| "calibrated_input_scales_preserved_exactly": true, | |
| "elapsed_seconds": 1.212, | |
| "emitted_expert_matrices": 0, | |
| "finite_positive_global_and_block_scales": true, | |
| "fused_source_tensors": 0, | |
| "gate_up_pairs_checked": 0, | |
| "output": "model-00131-of-00131.safetensors", | |
| "output_bytes": 1271398528, | |
| "output_keys": 1, | |
| "sample_reconstruction": {}, | |
| "schema": "blackfrost.cyber_frost_nvfp4.shared_w13.v1", | |
| "shard_number": 131, | |
| "source": "model-00131-of-00131.safetensors", | |
| "storage_mode": "copy_unchanged_cross_device" | |
| } | |
| ], | |
| "source": "/source", | |
| "source_keys": 1658, | |
| "source_shards": 131, | |
| "status": "complete", | |
| "storage_modes": { | |
| "copy_unchanged_cross_device": 35, | |
| "requantized_routed_experts_preserved_other_source_tensors": 96 | |
| } | |
| } | |