Mixtral_Alpace_v2

This model is a fine-tuned version of mistralai/Mixtral-8x7B-Instruct-v0.1 on the generator dataset. It achieves the following results on the evaluation set:

• Loss: 1.4981

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: 2.5e-05
• train_batch_size: 32
• eval_batch_size: 8
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_steps: 0.03
• training_steps: 200

Training results

Training Loss	Epoch	Step	Validation Loss
3.9673	0.83	10	3.6269
3.4669	1.67	20	3.0746
2.9203	2.5	30	2.5429
2.4531	3.33	40	2.2297
2.2242	4.17	50	2.1009
2.1178	5.0	60	2.0032
2.0242	5.83	70	1.9231
1.9462	6.67	80	1.8525
1.8682	7.5	90	1.7921
1.8218	8.33	100	1.7383
1.754	9.17	110	1.6900
1.7197	10.0	120	1.6468
1.672	10.83	130	1.6101
1.6248	11.67	140	1.5791
1.5936	12.5	150	1.5526
1.581	13.33	160	1.5324
1.5688	14.17	170	1.5168
1.5307	15.0	180	1.5066
1.532	15.83	190	1.5001
1.5254	16.67	200	1.4981

Framework versions

• PEFT 0.7.1
• Transformers 4.36.2
• Pytorch 2.1.2+cu121
• Datasets 2.16.1
• Tokenizers 0.15.0