Safetensors
qwen2
leave-one-out
loo-domain-knowledge
grpo
lr=5e-6
batch_size=2
group_size=2
max_steps=12
qlora
quantize=4bit_nf4
lora_rank=64
lora_alpha=128
lora_dropout=0.05
max_completion_length=2048
gradient_checkpointing
bf16
ddp_workers=2
hub_model_id=tinyllms/qwen2.5-7b-instruct-grpo-loo-domain-knowledge-20steps
ray_job=raysubmit_YbfK7VEEYMGjEp9h
- Xet hash:
- f8d5dc3430f301e3334c34268c690e5b8dfdc152b46f525db04e968ddb6379ba
- Size of remote file:
- 11.4 MB
- SHA256:
- ba0c439f7be467bf47d12a7e6f9adc6116201056fc60c67f431c679b7c16afc8
·
Xet efficiently stores Large Files inside Git, intelligently splitting files into unique chunks and accelerating uploads and downloads. More info.