WarpStateAI / model /sft-handoff.json
zerdovzad's picture
Upload 8 files
e5e74c7 verified
Raw History Blame Contribute Delete
5.62 kB
{
"archive_dir": "/workspace/WarpState/rtx6000_40b/work/archives/step-00076294",
"archive_sha256": "dc7605cecd6c807135e3d83e04496fa96079fef022bed1952c8cf21df5f573da",
"armed_at": "2026-09-14T09:43:24Z",
"command": [
"/workspace/WarpState/rtx6000_40b/.server-venv/bin/python",
"-B",
"-m",
"torch.distributed.run",
"--standalone",
"--nnodes",
"1",
"--nproc-per-node",
"2",
"--log-dir",
"/workspace/WarpState/rtx6000_40b/work/posttrain/step-00076294/logs/torchrun",
"/workspace/WarpState/rtx6000_40b/posttrain.py",
"--config",
"/workspace/WarpState/rtx6000_40b/configs/posttrain_10b.json",
"--work-dir",
"/workspace/WarpState/rtx6000_40b/work/posttrain/step-00076294/run",
"--data-dir",
"/workspace/WarpState/rtx6000_40b/work/posttrain/data",
"--weights",
"/workspace/WarpState/rtx6000_40b/work/posttrain/step-00076294/initialization/model.pt",
"--tokenizer",
"/workspace/WarpState/rtx6000_40b/tokenizer.model",
"--stop-file",
"/workspace/WarpState/rtx6000_40b/work/posttrain/step-00076294/control/training-stop-request"
],
"completion": {
"assistant_tokens_seen": 143695039,
"base_step": 76294,
"checkpoint": "/workspace/WarpState/rtx6000_40b/work/posttrain/step-00076294/run/latest.pt",
"completed": true,
"evaluation_file": "/workspace/WarpState/rtx6000_40b/work/posttrain/step-00076294/run/evaluation.json",
"format": "warpstate-posttrain-completion-v1",
"paused_for_evaluation": true,
"stage": "sft",
"step": 10312,
"timestamp": 1789419903.0140994,
"tokens_seen": 212321372,
"total_steps": 10312,
"val_loss": 1.1157395899572582
},
"data_ready": true,
"format": "warpstate-posttrain-handoff-v1",
"git_revision": "ea63c41cc541fdef8c36da8a85f7b625a2bcec40",
"message": "SFT finished; paused for evaluation. DPO has not been started.",
"phase": "complete",
"preflight": {
"accepted": true,
"bf16_compute": true,
"exact_resume_bookkeeping": true,
"final_validation_loss": 10.38550066947937,
"fixture_sha256": "bb6a842bbdc14a0a0b0537ba58e7630e16bbd09fc0613520303cff674a900dd3",
"format": "warpstate-posttrain-preflight-v1",
"fp32_parameters_and_adam": true,
"generator_sha256": "4489ae1f37c13e3226d6687147423a9f6a95a0509c135357291ea0a0a9d9664a",
"masked_tail": true,
"numerical_atol": 1e-07,
"numerical_comparison": {
"model": {
"bitwise_equal": true,
"max_absolute_difference": 0.0,
"tensor_count": 13
},
"optimizer": {
"bitwise_equal": true,
"max_absolute_difference": 0.0,
"tensor_count": 39
}
},
"numerical_rtol": 1e-05,
"optimizer_steps": 2,
"processes": 2,
"source_sha256": {
"ops/checkpoint_writer.py": "3c7fe494e8782802b93a439c05ddd6f416711ca1b3f0f8a196a97a48e37f0216",
"ops/loss_experiment.py": "12bf7822979692d5106e30544d0ed5d977fb06f556c068ef5f92ddf56ab8376e",
"ops/posttrain_data.py": "486d2e96e7e99ffd05caf0de7d92068ce47047d4641ede3515681cbb79a9ce10",
"ops/residual_experiment.py": "e82aa68e36c76a228156016eb04a02ce1a632a63ac6eda567ec9c4d258aae8f1",
"posttrain.py": "bd10e97e61b3289e83b7e79c76042bf7f04ce58f4af9067db97ac1fe60d4a411",
"run_40b.py": "9ad6533d781e376b3b40c16e6c29ef15063d39f917a93800489f17115be15ba0",
"warpstate/__init__.py": "f386c0880cb499d84159d8205667d250a72d0e8a102f0d7be6af0ba0ea33b22b",
"warpstate/config.py": "1d716b6c582bab21e8d5fcbaa6d9f3be4f792baf9b6abe826c72422f8eced6ad",
"warpstate/model.py": "025330ee58cca899695c7631e1a06d2efd006402c61e51106afe6fa2e1d35161"
},
"total_samples": 5
},
"source_sha256": {
"rtx6000_40b/chat.py": "148fd4f20d445ba01723d4c2625372132270884d5c44104f31a3cc13f7c02774",
"rtx6000_40b/configs/posttrain_10b.json": "a0f8d06e38f72245e046a6a888053c92aa23627256599348f5889886d61df179",
"rtx6000_40b/ops/checkpoint_writer.py": "3c7fe494e8782802b93a439c05ddd6f416711ca1b3f0f8a196a97a48e37f0216",
"rtx6000_40b/ops/launch_training.py": "fe9a204023300632766aabe7ff4b0cd22dba5ca7a306040da7ce3b21c9305e08",
"rtx6000_40b/ops/loss_experiment.py": "12bf7822979692d5106e30544d0ed5d977fb06f556c068ef5f92ddf56ab8376e",
"rtx6000_40b/ops/posttrain_data.py": "486d2e96e7e99ffd05caf0de7d92068ce47047d4641ede3515681cbb79a9ce10",
"rtx6000_40b/ops/posttrain_handoff.py": "8f7bb104183bb59802774c5500965d3ac2d1ffd1cb3de18256d075d5414f7c08",
"rtx6000_40b/ops/posttrain_preflight.py": "4489ae1f37c13e3226d6687147423a9f6a95a0509c135357291ea0a0a9d9664a",
"rtx6000_40b/ops/residual_experiment.py": "e82aa68e36c76a228156016eb04a02ce1a632a63ac6eda567ec9c4d258aae8f1",
"rtx6000_40b/ops/training-env.sh": "f5c9d6557d1dd9721e2e4c7bdeef001d994481c863c41bdd4080fd88484aa6f1",
"rtx6000_40b/posttrain.py": "bd10e97e61b3289e83b7e79c76042bf7f04ce58f4af9067db97ac1fe60d4a411",
"rtx6000_40b/run_40b.py": "9ad6533d781e376b3b40c16e6c29ef15063d39f917a93800489f17115be15ba0",
"rtx6000_40b/warpstate/__init__.py": "f386c0880cb499d84159d8205667d250a72d0e8a102f0d7be6af0ba0ea33b22b",
"rtx6000_40b/warpstate/config.py": "1d716b6c582bab21e8d5fcbaa6d9f3be4f792baf9b6abe826c72422f8eced6ad",
"rtx6000_40b/warpstate/model.py": "025330ee58cca899695c7631e1a06d2efd006402c61e51106afe6fa2e1d35161"
},
"target": {
"step": 76294,
"tokenizer_sha256": "ebae4b4310068ee102f7bf68ac68b79ee1425baa32de3ca8ff8138692529d84e",
"tokens_seen": 10000007168
},
"updated_at": "2026-09-14T21:05:23Z",
"work_dir": "/workspace/WarpState/rtx6000_40b/work/posttrain/step-00076294"
}