File size: 1,980 Bytes
9e05253 | 1 2 3 4 5 6 7 8 9 | # Timestamp: 2026-09-28T06:50:00.269218+00:00
# Source revisions: {"compressed-tensors": "a2a210e9001d6ac45b6f294f8a66a87384c7da55", "llm-compressor": "0349b68f4a7c5b0b535596df8e8d171911c8acee", "speculators": "fd69f5326d695e98434196b3860956494f4c0004"}
# Source status: {"compressed-tensors": "", "llm-compressor": "", "speculators": "M AGENTS.md\n M scripts/evaluate/evaluate.py\n M scripts/evaluate/perf_utils.py\n M scripts/launch_vllm.py\n M tests/e2e/smoke/test_finetuning_sanity.py\n M tests/unit/cli/test_cli.py\n M tests/unit/scripts/test_launch_vllm.py\n?? CONTEXT.md\n?? docs/agents/\n?? local/\n?? scripts/evaluate/point.py\n?? tests/unit/scripts/test_eval_point.py"}
# Package versions: {"compressed-tensors": "0.18.1.dev28+ga2a210e", "llmcompressor": "0.13.1.dev108+g0349b68f4", "torch": "2.13.0", "transformers": "5.16.1"}
# Source SHA-256: {"dquant/calibration.py": "f9dc82a58e1bc46ee944fc6a3e18b3b3d50361e89602d5ecf5ddd1e2fe63aa22", "dquant/quantize.py": "cb83ed43c02204bc820dfe06cf436b52e2c0616ac20f5224ba1869ab28375064", "scripts/30_quantize.sh": "053e6c9445bc1078352606b0f14ef32b9479d744655917c414d288c3b30142a6", "scripts/run_added_quantizations.sh": "4cd2e76cd9c8650c95d36640256c49df1167625c10cd7c798db7010e7c8a336d", "source_model/train_command.txt": "ffe1931dc6f16184ffb7faa5310ba05523a86319acbf4fc09ff5e04d0efe5505"}
# Environment: {"CUDA_VISIBLE_DEVICES": "0", "HF_HOME": "/data/fast/hf_cache", "OMP_NUM_THREADS": "8", "OPENBLAS_NUM_THREADS": "8"}
# Backfill note: Source copies captured after completed run. Operational FP8_BLOCK recipe is unchanged; later edits updated labels, verification, and provenance writing.
/workspace/speculators/.venv/bin/python -m dquant.quantize --scheme fp8_block --model /data/fast/hf_cache/hub/models--RedHatAI--Qwen3-8B-speculator.dflash/snapshots/1a11b170eb65c8a62c80ecd01dfe22a5907298e6 --output-dir /data/fast/drafter-quant/out/add-evals-20260928/fp8_block-datafree --calib-data none --device cuda:0 --max-workers 1
|