metadata

license: apache-2.0
base_model: mistralai/Mistral-7B-v0.1
tags:
  - generated_from_trainer
model-index:
  - name: Mistral_Sparse_refined_web_70p_2024-02-15
    results: []

Mistral_Sparse_refined_web_70p_2024-02-15

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: 2.2065

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-05
train_batch_size: 1
eval_batch_size: 1
seed: 0
distributed_type: multi-GPU
num_devices: 2
gradient_accumulation_steps: 8
total_train_batch_size: 16
total_eval_batch_size: 2
optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
lr_scheduler_type: linear
training_steps: 1000

Training results

Training Loss	Epoch	Step	Validation Loss
3.0372	0.0	25	3.1256
2.6176	0.01	50	2.8951
2.5321	0.01	75	2.7409
2.4603	0.02	100	2.6753
2.4033	0.02	125	2.6424
2.4821	0.02	150	2.6147
2.4008	0.03	175	2.5858
2.3651	0.03	200	2.5688
2.3873	0.04	225	2.5565
2.4145	0.04	250	2.5470
2.3295	0.04	275	2.5321
2.3458	0.05	300	2.5185
2.3587	0.05	325	2.5146
2.1873	0.06	350	2.5093
2.3502	0.06	375	2.5093
2.3837	0.06	400	2.5021
2.3747	0.07	425	2.4994
2.3292	0.07	450	2.4957
2.2438	0.08	475	2.4940
2.3102	0.08	500	2.4889
2.3791	0.08	525	2.4858
2.2743	0.09	550	2.4827
2.4148	0.09	575	2.4813
2.2115	0.1	600	2.4830
2.2963	0.1	625	2.4834
2.3762	0.1	650	2.4805
2.3657	0.11	675	2.4764
2.3219	0.11	700	2.4746
2.3166	0.12	725	2.4712
2.2193	0.12	750	2.4747
2.2629	0.12	775	2.4703
2.3504	0.13	800	2.4732
2.3523	0.13	825	2.4662
2.3362	0.14	850	2.4645
2.202	0.14	875	2.4659
2.2795	0.14	900	2.4682
2.2254	0.15	925	2.4621
2.3507	0.15	950	2.4642
2.2825	0.16	975	2.4624
2.3301	0.16	1000	2.4603

Framework versions

Transformers 4.36.2
Pytorch 2.1.2+cu121
Datasets 2.15.0
Tokenizers 0.15.0