You can deploy this model and start earning money today!

RichardErkhov/
thrunlab_-_Mistral_Sparse_refined_web_relu_2024-03-11-4bits

Name: thrunlab_-_Mistral_Sparse_refined_web_relu_2024-03-11-4bits
Availability: InStock

safetensors

Views

No views yet

Quantization made by Richard Erkhov.

Github

Discord

Request more models

Mistral_Sparse_refined_web_relu_2024-03-11 - bnb 4bits

Model creator: https://huggingface.co/thrunlab/
Original model: https://huggingface.co/thrunlab/Mistral_Sparse_refined_web_relu_2024-03-11/

Original model description:

license: apache-2.0 base_model: mistralai/Mistral-7B-v0.1 tags:

generated_from_trainer model-index:
name: Mistral_Sparse_refined_web_relu_2024-03-11 results: []

Mistral_Sparse_refined_web_relu_2024-03-11

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: 2.3749

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: 2600

Training results

Training Loss	Epoch	Step	Validation Loss
8.7876	0.0	25	8.7120
8.1837	0.01	50	8.1741
7.7529	0.01	75	7.8037
7.5336	0.02	100	7.6036
7.1718	0.02	125	7.1058
4.7663	0.02	150	4.7163
3.7548	0.03	175	3.8001
3.2625	0.03	200	3.4192
3.0886	0.04	225	3.2355
2.9928	0.04	250	3.1153
2.8555	0.04	275	3.0392
2.8289	0.05	300	2.9870
2.7933	0.05	325	2.9505
2.6073	0.06	350	2.9160
2.7488	0.06	375	2.8941
2.7795	0.06	400	2.8722
2.7317	0.07	425	2.8537
2.6982	0.07	450	2.8407
2.5823	0.08	475	2.8258
2.6419	0.08	500	2.8164
2.7195	0.08	525	2.8082
2.6239	0.09	550	2.7980
2.7273	0.09	575	2.7869
2.5436	0.1	600	2.7809
2.6159	0.1	625	2.7761
2.6563	0.1	650	2.7666
2.6728	0.11	675	2.7573
2.6047	0.11	700	2.7509
2.6237	0.12	725	2.7493
2.5305	0.12	750	2.7458
2.5329	0.12	775	2.7392
2.6538	0.13	800	2.7359
2.6076	0.13	825	2.7310
2.5928	0.14	850	2.7279
2.455	0.14	875	2.7246
2.5579	0.14	900	2.7252
2.4948	0.15	925	2.7194
2.6219	0.15	950	2.7181
2.5387	0.16	975	2.7139
2.5734	0.16	1000	2.7134
2.6012	0.16	1025	2.7115
2.63	0.17	1050	2.7076
2.6361	0.17	1075	2.7045
2.5534	0.18	1100	2.7046
2.5756	0.18	1125	2.7031
2.5632	0.18	1150	2.6989
2.5971	0.19	1175	2.6960
2.4719	0.19	1200	2.6963
2.5377	0.2	1225	2.6944
2.552	0.2	1250	2.6907
2.5748	0.2	1275	2.6894
2.5799	0.21	1300	2.6877
2.5569	0.21	1325	2.6834
2.4413	0.22	1350	2.6822
2.5232	0.22	1375	2.6822
2.5346	0.22	1400	2.6806
2.479	0.23	1425	2.6791
2.4585	0.23	1450	2.6803
2.4104	0.24	1475	2.6776
2.4961	0.24	1500	2.6792
2.4219	0.24	1525	2.6770
2.4658	0.25	1550	2.6736
2.5875	0.25	1575	2.6755
2.5376	0.26	1600	2.6705
2.5466	0.26	1625	2.6726
2.5889	0.26	1650	2.6704
2.4973	0.27	1675	2.6667
2.5409	0.27	1700	2.6681
2.5386	0.28	1725	2.6658
2.5234	0.28	1750	2.6666
2.5066	0.28	1775	2.6619
2.4283	0.29	1800	2.6629
2.5253	0.29	1825	2.6623
2.5179	0.3	1850	2.6599
2.5023	0.3	1875	2.6608
2.5253	0.3	1900	2.6602
2.5788	0.31	1925	2.6602
2.5307	0.31	1950	2.6596
2.5108	0.32	1975	2.6593
2.462	0.32	2000	2.6597
2.5028	0.32	2025	2.6577
2.48	0.33	2050	2.6538
2.4742	0.33	2075	2.6534
2.554	0.34	2100	2.6544
2.5987	0.34	2125	2.6547
2.5224	0.34	2150	2.6550
2.4526	0.35	2175	2.6510
2.503	0.35	2200	2.6484
2.4648	0.36	2225	2.6487
2.4568	0.36	2250	2.6481
2.5701	0.36	2275	2.6465
2.5403	0.37	2300	2.6467
2.435	0.37	2325	2.6472
2.4823	0.38	2350	2.6479
2.536	0.38	2375	2.6468
2.5171	0.38	2400	2.6470
2.3852	0.39	2425	2.6475
2.3807	0.39	2450	2.6471
2.4753	0.4	2475	2.6456
2.5507	0.4	2500	2.6442
2.5331	0.4	2525	2.6441
2.3988	0.41	2550	2.6415
2.425	0.41	2575	2.6403
2.5062	0.42	2600	2.6429

Framework versions

Transformers 4.36.2
Pytorch 2.1.2+cu121
Datasets 2.15.0
Tokenizers 0.15.0

RichardErkhov/thrunlab_-_Mistral_Sparse_refined_web_relu_2024-03-11-4bits

Original model description:

Mistral_Sparse_refined_web_relu_2024-03-11

Model description

Intended uses & limitations

Training and evaluation data

Training procedure

Training hyperparameters

Training results

Framework versions

RichardErkhov/
thrunlab_-_Mistral_Sparse_refined_web_relu_2024-03-11-4bits