Mistral_Sparse_refined_web_50p_cut_pre_mlp_cut_pre_attn_2024-03-23

thrunlab/Mistral_Sparse_refined_web_50p_cut_pre_mlp_cut_pre_attn_2024-03-23

Mistral_Sparse_refined_web_50p_cut_pre_mlp_cut_pre_attn_2024-03-23 is a text generation model from thrunlab. Use it when you need the model to write or continue text. It is set up for transformers. The card lists the license as apache-2.0.

should probably proofread and complete it, then remove this comment. --

Downloads · 30 days

11% of all-time downloads

All-time downloads

Public

Parameters

7.2B

16.7 GB on disk

Likes

Public

Hugging Face

Repo makeup

Click a slice to open those files.

.safetensors14.8 GB · 100%

At a glance

Task: Text Generation
Library: transformers
License: apache-2.0
Model type: sparse_mistral
Access: Public
Created: Mar 23, 2024
Updated: Mar 23, 2024
SHA: ca5c97a1

Try a prompt

Base models

mistralai/Mistral-7B-v0.1

Model card

From the Hugging Face model README

Mistral_Sparse_refined_web_50p_cut_pre_mlp_cut_pre_attn_2024-03-23

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: 2.1772

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: 4
gradient_accumulation_steps: 4
total_train_batch_size: 16
total_eval_batch_size: 4
optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
lr_scheduler_type: linear
training_steps: 2000

Training results

Training Loss	Epoch	Step	Validation Loss
2.5343	0.0	25	2.6759
2.3148	0.01	50	2.6092
2.356	0.01	75	2.5789
2.32	0.02	100	2.5624
2.3107	0.02	125	2.5371
2.3921	0.02	150	2.5225
2.3138	0.03	175	2.5241
2.3012	0.03	200	2.5020
2.3375	0.04	225	2.5107
2.3714	0.04	250	2.5091
2.2891	0.04	275	2.4997
2.2997	0.05	300	2.4911
2.3221	0.05	325	2.4886
2.1478	0.06	350	2.4888
2.3131	0.06	375	2.4837
2.3461	0.06	400	2.4811
2.3443	0.07	425	2.4747
2.3041	0.07	450	2.4760
2.2192	0.08	475	2.4765
2.2829	0.08	500	2.4698
2.347	0.08	525	2.4709
2.2503	0.09	550	2.4663
2.3858	0.09	575	2.4688
2.1889	0.1	600	2.4642
2.2762	0.1	625	2.4636
2.355	0.1	650	2.4612
2.3422	0.11	675	2.4697
2.304	0.11	700	2.4545
2.2965	0.12	725	2.4606
2.2014	0.12	750	2.4610
2.2404	0.12	775	2.4558
2.3355	0.13	800	2.4527
2.3421	0.13	825	2.4503
2.3193	0.14	850	2.4534
2.1828	0.14	875	2.4556
2.2652	0.14	900	2.4490
2.203	0.15	925	2.4531
2.3358	0.15	950	2.4626
2.2625	0.16	975	2.4448
2.3168	0.16	1000	2.4459
2.3163	0.16	1025	2.4438
2.3319	0.17	1050	2.4427
2.36	0.17	1075	2.4439
2.2884	0.18	1100	2.4501
2.3153	0.18	1125	2.4493
2.2807	0.18	1150	2.4386
2.3341	0.19	1175	2.4484
2.1909	0.19	1200	2.4458
2.2831	0.2	1225	2.4417
2.2759	0.2	1250	2.4472
2.3158	0.2	1275	2.4422
2.3413	0.21	1300	2.4450
2.3078	0.21	1325	2.4494
2.2061	0.22	1350	2.4451
2.2846	0.22	1375	2.4359
2.2929	0.22	1400	2.4358
2.2341	0.23	1425	2.4389
2.2222	0.23	1450	2.4452
2.1849	0.24	1475	2.4427
2.2468	0.24	1500	2.4396
2.1769	0.24	1525	2.4431
2.2323	0.25	1550	2.4403
2.3575	0.25	1575	2.4421
2.3032	0.26	1600	2.4437
2.2787	0.26	1625	2.4390
2.3523	0.26	1650	2.4374
2.2613	0.27	1675	2.4397
2.3048	0.27	1700	2.4300
2.3016	0.28	1725	2.4377
2.2821	0.28	1750	2.4394
2.2642	0.28	1775	2.4356
2.2181	0.29	1800	2.4369
2.2917	0.29	1825	2.4429
2.2922	0.3	1850	2.4364
2.2718	0.3	1875	2.4446
2.2961	0.3	1900	2.4378
2.3482	0.31	1925	2.4374
2.2985	0.31	1950	2.4369
2.3086	0.32	1975	2.4352
2.2412	0.32	2000	2.4395

Framework versions

Transformers 4.36.2
Pytorch 2.1.2+cu121
Datasets 2.15.0
Tokenizers 0.15.0

Task: Text Generation
Library: transformers
Type: sparse_mistral
License: apache-2.0
Created: Mar 23, 2024
Updated: Mar 23, 2024