DPO-PairRM-5-SMI-lr-1e6-iteration-5-t-7e-beta-15e3-1-iteration-6e1-confidence
This model is a fine-tuned version of mistralai/Mistral-7B-Instruct-v0.2 on the None dataset. It achieves the following results on the evaluation set:
- Loss: 0.6363
- Rewards/chosen: -2.1229
- Rewards/rejected: -2.5297
- Rewards/accuracies: 0.6588
- Rewards/margins: 0.4068
- Rewards/mix Margin: 0.1426
- Logps/rejected: -446.6819
- Logps/chosen: -434.8683
- Logits/rejected: -2.0316
- Logits/chosen: -2.0446
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: 1e-06
- train_batch_size: 1
- eval_batch_size: 1
- seed: 42
- distributed_type: multi-GPU
- num_devices: 4
- gradient_accumulation_steps: 16
- total_train_batch_size: 64
- total_eval_batch_size: 4
- 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
Framework versions
- Transformers 4.36.2
- Pytorch 2.1.2
- Datasets 2.17.1
- Tokenizers 0.15.1
- Downloads last month
- 3
Model is too large to load in Inference API (serverless). To try the model, launch it on Inference Endpoints (dedicated)
instead.