thrunlab_-_sparse_mistral_7b_refined_web_50p_2024-04-13-4bits

RichardErkhov/thrunlab_-_sparse_mistral_7b_refined_web_50p_2024-04-13-4bits

thrunlab_-_sparse_mistral_7b_refined_web_50p_2024-04-13-4bits is a machine learning model from RichardErkhov. Use it for the machine learning task on the model card, and read the license before you ship it in a product.

Downloads · 30 days

Access

Public

Updated Apr 14, 2024

Repo size

281 MB

Likes

Public

Hugging Face

Repo makeup

Click a slice to open those files.

.safetensors281 MB · 99%

At a glance

Access: Public
Created: Apr 14, 2024
Updated: Apr 14, 2024
SHA: 89108177

Model card

From the Hugging Face model README

Quantization made by Richard Erkhov.

Github

Discord

Request more models

sparse_mistral_7b_refined_web_50p_2024-04-13 - bnb 4bits

Model creator: https://huggingface.co/thrunlab/
Original model: https://huggingface.co/thrunlab/sparse_mistral_7b_refined_web_50p_2024-04-13/

Original model description:

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

generated_from_trainer model-index:
name: sparse_mistral_7b_refined_web_50p_2024-04-13 results: []

sparse_mistral_7b_refined_web_50p_2024-04-13

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: 2.1985

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: 4
seed: 0
distributed_type: multi-GPU
num_devices: 4
gradient_accumulation_steps: 8
total_train_batch_size: 32
total_eval_batch_size: 16
optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
lr_scheduler_type: linear
training_steps: 2350

Training results

Training Loss	Epoch	Step	Validation Loss
2.3391	0.01	25	2.4196
2.2711	0.02	50	2.3577
2.3054	0.02	75	2.3158
2.2795	0.03	100	2.2966
2.3175	0.04	125	2.2846
2.2388	0.05	150	2.2766
2.1679	0.06	175	2.2705
2.2996	0.06	200	2.2678
2.2788	0.07	225	2.2647
2.2448	0.08	250	2.2637
2.1837	0.09	275	2.2624
2.2089	0.1	300	2.2621
2.2686	0.1	325	2.2601
2.2254	0.11	350	2.2593
2.162	0.12	375	2.2590
2.2687	0.13	400	2.2563
2.2595	0.14	425	2.2571
2.186	0.14	450	2.2564
2.2689	0.15	475	2.2580
2.2472	0.16	500	2.2554
2.2005	0.17	525	2.2553
2.1983	0.18	550	2.2552
2.2388	0.18	575	2.2547
2.1443	0.19	600	2.2555
2.2198	0.2	625	2.2534
2.3008	0.21	650	2.2536
2.179	0.22	675	2.2521
2.2069	0.22	700	2.2531
2.1819	0.23	725	2.2526
2.1218	0.24	750	2.2536
2.1845	0.25	775	2.2515
2.2167	0.26	800	2.2510
2.2252	0.26	825	2.2520
2.1664	0.27	850	2.2519
2.1853	0.28	875	2.2530
2.1499	0.29	900	2.2513
2.2763	0.3	925	2.2517
2.2528	0.3	950	2.2518
2.2505	0.31	975	2.2500
2.1683	0.32	1000	2.2502
2.2177	0.33	1025	2.2501
2.238	0.34	1050	2.2516
2.193	0.34	1075	2.2507
2.2025	0.35	1100	2.2502
2.0944	0.36	1125	2.2512
2.2272	0.37	1150	2.2508
2.2264	0.38	1175	2.2500
2.1837	0.38	1200	2.2507
2.1444	0.39	1225	2.2489
2.2464	0.4	1250	2.2499
2.1388	0.41	1275	2.2508
2.193	0.42	1300	2.2492
2.2376	0.42	1325	2.2506
2.2212	0.43	1350	2.2478
2.2002	0.44	1375	2.2488
2.2729	0.45	1400	2.2484
2.2329	0.46	1425	2.2473
2.1919	0.46	1450	2.2481
2.2102	0.47	1475	2.2475
2.1466	0.48	1500	2.2473
2.1819	0.49	1525	2.2478
2.2558	0.5	1550	2.2468
2.2137	0.5	1575	2.2463
2.2288	0.51	1600	2.2466
2.1479	0.52	1625	2.2468
2.1726	0.53	1650	2.2471
2.1805	0.54	1675	2.2454
2.1505	0.54	1700	2.2470
2.1337	0.55	1725	2.2465
2.2413	0.56	1750	2.2460
2.152	0.57	1775	2.2478
2.2669	0.58	1800	2.2471
2.2925	0.58	1825	2.2465
2.222	0.59	1850	2.2457
2.1308	0.6	1875	2.2466
2.201	0.61	1900	2.2456
2.2247	0.62	1925	2.2460
2.2426	0.62	1950	2.2463
2.2312	0.63	1975	2.2465
2.2679	0.64	2000	2.2464
2.1928	0.65	2025	2.2463
2.2087	0.66	2050	2.2455
2.1792	0.66	2075	2.2470
2.252	0.67	2100	2.2468
2.2018	0.68	2125	2.2456
2.2006	0.69	2150	2.2451
2.2076	0.7	2175	2.2449
2.2436	0.7	2200	2.2460
2.2156	0.71	2225	2.2477
2.1348	0.72	2250	2.2455
2.1338	0.73	2275	2.2450
2.2147	0.74	2300	2.2455
2.2766	0.74	2325	2.2444
2.204	0.75	2350	2.2458

Framework versions

Transformers 4.36.2
Pytorch 2.1.2+cu121
Datasets 2.15.0
Tokenizers 0.15.0

Created: Apr 14, 2024
Updated: Apr 14, 2024