-
Add HuggingFace Hub checkpoint persistence - upload and download checkpoints between jobs
renpas22
-
Restore SFT to 10000 steps - will resume from checkpoint at 7000
renpas22
-
Reduce training steps to fit within job timeout (SFT: 5k, PRM: 2k, RL: 3k)
renpas22
-
Quote mixed_precision value
renpas22
-
Add inference config parameters
renpas22
-
Add missing RL/PPO config parameters
renpas22
-
Fix model name to use valid Qwen2-VL model
renpas22
-
Enable 8-bit quantization and reduce batch size for memory
renpas22
-
Add required top-level config keys for trainer
renpas22
-
Add training scripts and configs
renpas22