quyanh
/

pythia-2.8b-dpo

Generated from Trainer

Model card Files Files and versions

pythia-2.8b-dpo

87.5 MB

1 contributor

History: 103 commits

quyanh's picture

DPO LoRA checkpoint - step 10000 (eval_loss: 0.6355)

544a0c8 verified 3 months ago

.gitattributes

1.52 kB

initial commit 3 months ago
README.md

2.52 kB

Training in progress, step 100 3 months ago
adapter_config.json

941 Bytes

DPO LoRA checkpoint - step 100 (eval_loss: 0.6984) 3 months ago
adapter_model.safetensors

83.9 MB
xet

DPO LoRA checkpoint - step 10000 (eval_loss: 0.6355) 3 months ago
special_tokens_map.json

587 Bytes

Tokenizer - DPO LoRA checkpoint - step 100 (eval_loss: 0.6984) 3 months ago
tokenizer.json

3.56 MB

Tokenizer - DPO LoRA checkpoint - step 100 (eval_loss: 0.6984) 3 months ago
tokenizer_config.json

5.07 kB

Tokenizer - DPO LoRA checkpoint - step 100 (eval_loss: 0.6984) 3 months ago
training_args.bin
Detected Pickle imports (11)
- "__main__.DPOTrainingArguments",
- "accelerate.state.PartialState",
- "transformers.trainer_utils.IntervalStrategy",
- "transformers.trainer_utils.SchedulerType",
- "trl.trainer.dpo_config.FDivergenceType",
- "transformers.trainer_utils.HubStrategy",
- "transformers.trainer_utils.SaveStrategy",
- "transformers.training_args.OptimizerNames",
- "torch.device",
- "transformers.trainer_pt_utils.AcceleratorConfig",
- "accelerate.utils.dataclasses.DistributedType"
How to fix it?
6.8 kB
xet

Training in progress, step 100 3 months ago