Download runs/launch-records/llama3b-code-ordered.json from RL-Forgetting-Experiments-3/code-sft-baselines-20261027-artifacts: direct link, hf CLI and curl.
- Browser
- Download file 797 Bytes
-
https://huggingface.co/datasets/RL-Forgetting-Experiments-3/code-sft-baselines-20261027-artifacts/resolve/main/runs/launch-records/llama3b-code-ordered.json
- Command line
-
hf download hf://datasets/RL-Forgetting-Experiments-3/code-sft-baselines-20261027-artifacts/runs/launch-records/llama3b-code-ordered.json
-
curl -L -o llama3b-code-ordered.json https://huggingface.co/datasets/RL-Forgetting-Experiments-3/code-sft-baselines-20261027-artifacts/resolve/main/runs/launch-records/llama3b-code-ordered.json
797 Bytes
| { | |
| "arm": { | |
| "name": "llama3b-code-ordered", | |
| "model": "llama3b", | |
| "dataset": "llama3b_code_sft_data", | |
| "base": "Llama-3.2-3B-Instruct", | |
| "order": "ordered", | |
| "strategy": "none" | |
| }, | |
| "source_revision": "ec5a459cf6def8d1cf420b10a6b9ac5359681d9f", | |
| "checkpoint_fix": "recovery-20261020", | |
| "environment": { | |
| "MODEL_NAME": "llama3b", | |
| "DOMAIN": "code", | |
| "DATASET": "llama3b_code_sft_data", | |
| "BASE_MODEL": "/mnt/mystorageoutput/projects/rl-forgetting/base_models/Llama-3.2-3B-Instruct", | |
| "LRS": "1e-5", | |
| "ORDERS": "ordered", | |
| "REPLAY_STRATEGY": "none", | |
| "REPLAY_LAMBDA": "0", | |
| "NPROC": "4", | |
| "MICRO_BATCH_SIZE": "1", | |
| "EFFECTIVE_BATCH_SIZE": "256", | |
| "NUM_EPOCHS": "1", | |
| "SEED": "42", | |
| "MAX_SEQ_LENGTH": "4096", | |
| "SKIP_EVAL": "1" | |
| } | |
| } |