imp_sam_iter_3
This model is a fine-tuned version of YYYYYYibo/imp_sam_iter_2 on the updated and the original datasets. It achieves the following results on the evaluation set:
- Loss: 0.6669
- Rewards/chosen: -0.0263
- Rewards/rejected: -0.0862
- Rewards/accuracies: 0.6440
- Rewards/margins: 0.0600
- Logps/rejected: -269.8819
- Logps/chosen: -289.5703
- Logits/rejected: -2.5401
- Logits/chosen: -2.6276
Model description
More information needed
Intended uses & limitations
More information needed
Training and evaluation data
More information needed
Training procedure
Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 5e-06
- train_batch_size: 2
- eval_batch_size: 2
- seed: 42
- distributed_type: multi-GPU
- num_devices: 4
- gradient_accumulation_steps: 16
- total_train_batch_size: 128
- total_eval_batch_size: 8
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: cosine
- lr_scheduler_warmup_ratio: 0.1
- num_epochs: 1
Training results
Training Loss | Epoch | Step | Validation Loss | Rewards/chosen | Rewards/rejected | Rewards/accuracies | Rewards/margins | Logps/rejected | Logps/chosen | Logits/rejected | Logits/chosen |
---|---|---|---|---|---|---|---|---|---|---|---|
0.6781 | 0.65 | 100 | 0.6669 | -0.0263 | -0.0862 | 0.6440 | 0.0600 | -269.8819 | -289.5703 | -2.5401 | -2.6276 |
Framework versions
- PEFT 0.7.1
- Transformers 4.36.2
- Pytorch 2.1.2+cu121
- Datasets 2.14.6
- Tokenizers 0.15.2
- Downloads last month
- 2
Model tree for YYYYYYibo/imp_sam_iter_3
Base model
mistralai/Mistral-7B-v0.1
Finetuned
alignment-handbook/zephyr-7b-sft-full
Adapter
this model