Summary4500_M2_1000steps_1e8rate_05beta_CSFTDPO

tsavage68/Summary4500_M2_1000steps_1e8rate_05beta_CSFTDPO

Summary4500_M2_1000steps_1e8rate_05beta_CSFTDPO is a text generation model from tsavage68. 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

19% of all-time downloads

All-time downloads

Public

Parameters

7.2B

14.5 GB on disk

Likes

Public

Hugging Face

Repo makeup

Click a slice to open those files.

.safetensors29 GB · 100%

At a glance

Task: Text Generation
Library: transformers
License: apache-2.0
Model type: mistral
Access: Public
Created: Jul 17, 2024
Updated: Jul 17, 2024
SHA: df0f465c

Try a prompt

Base models

tsavage68/Summary4500_M2_200steps_1e7rate_SFT

Model card

From the Hugging Face model README

Hyponatremia_M2_1000steps_1e8rate_05beta_CSFTDPO

This model is a fine-tuned version of tsavage68/Summary4500_M2_200steps_1e7rate_SFT on an unknown dataset. It achieves the following results on the evaluation set:

Loss: 0.6072
Rewards/chosen: -0.0220
Rewards/rejected: -0.2274
Rewards/accuracies: 0.7680
Rewards/margins: 0.2055
Logps/rejected: -153.1847
Logps/chosen: -93.7837
Logits/rejected: -2.3491
Logits/chosen: -2.3018

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-08
train_batch_size: 1
eval_batch_size: 1
seed: 42
optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
lr_scheduler_type: cosine
lr_scheduler_warmup_steps: 100
training_steps: 1000

Training results

Training Loss	Epoch	Step	Validation Loss	Rewards/chosen	Rewards/rejected	Rewards/accuracies	Rewards/margins	Logps/rejected	Logps/chosen	Logits/rejected	Logits/chosen
0.7203	0.0112	50	0.6987	-0.0033	-0.0121	0.5140	0.0089	-152.7540	-93.7463	-2.3525	-2.3051
0.7141	0.0224	100	0.7007	-0.0065	-0.0124	0.5120	0.0059	-152.7545	-93.7527	-2.3518	-2.3044
0.7206	0.0336	150	0.6687	-0.0031	-0.0746	0.5660	0.0715	-152.8790	-93.7460	-2.3524	-2.3049
0.587	0.0448	200	0.6498	-0.0119	-0.1255	0.6020	0.1137	-152.9809	-93.7635	-2.3514	-2.3040
0.6645	0.0559	250	0.6356	0.0008	-0.1439	0.6940	0.1447	-153.0175	-93.7381	-2.3513	-2.3039
0.6374	0.0671	300	0.6231	-0.0060	-0.1737	0.7240	0.1676	-153.0771	-93.7519	-2.3518	-2.3044
0.5996	0.0783	350	0.6137	-0.0051	-0.1962	0.7200	0.1911	-153.1222	-93.7499	-2.3499	-2.3025
0.4867	0.0895	400	0.6126	-0.0077	-0.1993	0.7540	0.1916	-153.1284	-93.7552	-2.3512	-2.3037
0.5363	0.1007	450	0.6146	-0.0011	-0.1887	0.7420	0.1876	-153.1071	-93.7419	-2.3512	-2.3038
0.6204	0.1119	500	0.6110	-0.0072	-0.2069	0.7460	0.1998	-153.1436	-93.7541	-2.3507	-2.3033
0.6675	0.1231	550	0.6090	-0.0054	-0.2048	0.7420	0.1994	-153.1393	-93.7506	-2.3507	-2.3032
0.6008	0.1343	600	0.6026	-0.0041	-0.2203	0.7640	0.2162	-153.1705	-93.7480	-2.3498	-2.3024
0.5632	0.1454	650	0.6022	0.0008	-0.2179	0.7560	0.2187	-153.1656	-93.7381	-2.3506	-2.3032
0.6104	0.1566	700	0.6113	-0.0078	-0.2032	0.7540	0.1955	-153.1363	-93.7553	-2.3516	-2.3042
0.6059	0.1678	750	0.6027	-0.0019	-0.2188	0.7760	0.2169	-153.1674	-93.7435	-2.3507	-2.3033
0.5551	0.1790	800	0.5949	-0.0079	-0.2395	0.7640	0.2315	-153.2087	-93.7556	-2.3495	-2.3021
0.579	0.1902	850	0.6054	-0.0160	-0.2255	0.7720	0.2095	-153.1809	-93.7718	-2.3491	-2.3018
0.6122	0.2014	900	0.6072	-0.0220	-0.2274	0.7680	0.2055	-153.1847	-93.7837	-2.3491	-2.3018
0.5744	0.2126	950	0.6072	-0.0220	-0.2274	0.7680	0.2055	-153.1847	-93.7837	-2.3491	-2.3018
0.6718	0.2238	1000	0.6072	-0.0220	-0.2274	0.7680	0.2055	-153.1847	-93.7837	-2.3491	-2.3018

Framework versions

Transformers 4.42.4
Pytorch 2.0.0+cu117
Datasets 2.20.0
Tokenizers 0.19.1

Task: Text Generation
Library: transformers
Type: mistral
License: apache-2.0
Created: Jul 17, 2024
Updated: Jul 17, 2024