license, library_name, pipeline_tag, tags

license	library_name	pipeline_tag	tags
apache-2.0	transformers	text-generation	qwen3 reinforcement-learning rlsd

qwen3-8b-material-rlsd-ema005

Fine-tuned from Qwen/Qwen3-8B with RLSD (EMA 0.05) on the material split.

Validation Performance

Metric: val-aux/sciknoweval/reward/mean@16 from 10-step validation logs.

best mean@16	best step	final mean@16	final step
79.06%	100	79.06%	100

step	mean@16
10	69.95%
20	72.54%
30	75.40%
40	76.33%
50	76.13%
60	73.40%
70	74.60%
80	74.80%
90	76.13%
100	79.06%

Files included with this repo:

• metrics.json: parsed validation summary
• eval_mean16.csv: step-level validation curve data
• eval_mean16.png: validation curve plot

Important: the uploaded weights are the final global_step_100/actor checkpoint. If best step is earlier than 100, the best validation point is reported for tracking, but the corresponding actor weights may not be retained locally.

Checkpoint source: /mnt/mole/SDPO/L2T/checkpoints/datasets/sciknoweval/material/qwen3gen-material-RLSD-Qwen-Qwen3-8B-mbs8-decay0-ema0.05-train64-rollout8-lr1e-6-vllm0.8

W&B run: run-20260701_023114-1704cjhy

This upload uses global_step_100/actor converted from VERL FSDP shards to Hugging Face format.