metadata

license: apache-2.0
library_name: peft
tags:
  - generated_from_trainer
base_model: mistralai/Mistral-7B-v0.1
model-index:
  - name: Mistral-7B-v0.1_caselaw
    results: []

Mistral-7B-v0.1_caselaw

This model is a fine-tuned version of mistralai/Mistral-7B-v0.1 on the None dataset. It achieves the following results on the evaluation set:

Loss: 1.1640

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: 3e-05
train_batch_size: 4
eval_batch_size: 4
seed: 42
distributed_type: multi-GPU
num_devices: 4
gradient_accumulation_steps: 4
total_train_batch_size: 64
total_eval_batch_size: 16
optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
lr_scheduler_type: constant
lr_scheduler_warmup_ratio: 0.03
num_epochs: 2.0

Training results

Training Loss	Epoch	Step	Validation Loss
1.2324	0.07	50	1.2373
1.2114	0.13	100	1.2199
1.1831	0.2	150	1.2111
1.2027	0.26	200	1.2048
1.1827	0.33	250	1.2001
1.1696	0.39	300	1.1973
1.2186	0.46	350	1.1938
1.1795	0.52	400	1.1919
1.2167	0.59	450	1.1884
1.1992	0.66	500	1.1840
1.2032	0.72	550	1.1824
1.1841	0.79	600	1.1798
1.166	0.85	650	1.1789
1.1641	0.92	700	1.1761
1.1859	0.98	750	1.1752
1.132	1.05	800	1.1736
1.1461	1.12	850	1.1724
1.0965	1.18	900	1.1726
1.1064	1.25	950	1.1724
1.123	1.31	1000	1.1729
1.1079	1.38	1050	1.1695
1.12	1.44	1100	1.1707
1.1288	1.51	1150	1.1693
1.133	1.57	1200	1.1676
1.1647	1.64	1250	1.1693
1.1269	1.71	1300	1.1658
1.1332	1.77	1350	1.1657
1.1276	1.84	1400	1.1681
1.1361	1.9	1450	1.1633
1.1205	1.97	1500	1.1640

Framework versions

PEFT 0.7.1
Transformers 4.37.2
Pytorch 2.1.2+cu121
Datasets 2.17.1
Tokenizers 0.15.1