楚毓 c0d2c008b0 fix: bound LLM inference memory for batch screening

2026-09-23 10:11:41 +08:00

..

encoders

Add MoleculeSTM/MolE/RF/KNN baselines: encoders, scripts, and CV results

2026-07-28 14:14:10 +00:00

layers

fix: bound LLM inference memory for batch screening

2026-09-23 10:11:41 +08:00

__init__.py

Initial commit

2026-01-18 21:50:21 +08:00

benchmark.py

Add full ablation results and CheMeleon/UniMol encoders

2026-07-24 17:56:49 +00:00

final_train_optuna_cv.py

feat: 支持固定超参训练并部署 4 专家稀疏 MoE 模型

2026-08-15 12:22:21 +00:00

heads.py

feat: final_train_optuna_cv 支持 MoE + Qwen2.5-7B QLoRA + soft-RAG，产出可部署 checkpoint

2026-08-13 10:05:00 +00:00

models.py

fix: 修正 size 预测的反归一化并将 MoE+LLM 模型接入应用

2026-08-14 11:24:02 +00:00

nested_cv_optuna.py

fix: 修正 size 预测的反归一化并将 MoE+LLM 模型接入应用

2026-08-14 11:24:02 +00:00

predict.py

fix: bound LLM inference memory for batch screening

2026-09-23 10:11:41 +08:00

pretrain.py

Add full ablation results and CheMeleon/UniMol encoders

2026-07-24 17:56:49 +00:00

retrieval.py

feat: Qwen SoftRAG（RAG prompt/soft token/QLoRA）+ Morgan 检索旁路与实验结果

2026-07-07 17:47:34 +00:00

trainer_balanced.py

feat: final_train_optuna_cv 支持 MoE + Qwen2.5-7B QLoRA + soft-RAG，产出可部署 checkpoint

2026-08-13 10:05:00 +00:00

trainer.py

Add Mixture-of-Experts (MoE) layer support across modeling pipeline

2026-05-26 15:26:18 +08:00

visualization.py

Add loss visualization

2026-01-23 13:40:22 +08:00