IE_M2_1000steps_1e5rate_01beta_cSFTDPO

tsavage68/IE_M2_1000steps_1e5rate_01beta_cSFTDPO

IE_M2_1000steps_1e5rate_01beta_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

4% of all-time downloads

All-time downloads

131

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: Sep 23, 2024
Updated: Sep 23, 2024
SHA: 8549b65a

Try a prompt

Base models

tsavage68/IE_M2_1000steps_1e7rate_SFT

Model card

From the Hugging Face model README

IE_M2_1000steps_1e5rate_01beta_cSFTDPO

This model is a fine-tuned version of tsavage68/IE_M2_1000steps_1e7rate_SFT on an unknown dataset. It achieves the following results on the evaluation set:

Loss: 0.3743
Rewards/chosen: -2.2787
Rewards/rejected: -22.2937
Rewards/accuracies: 0.4600
Rewards/margins: 20.0150
Logps/rejected: -263.9586
Logps/chosen: -64.9924
Logits/rejected: -2.7884
Logits/chosen: -2.7491

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: 2
eval_batch_size: 1
seed: 42
gradient_accumulation_steps: 2
total_train_batch_size: 4
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
0.4505	0.4	50	0.3743	-2.1144	-14.8883	0.4600	12.7739	-189.9051	-63.3497	-2.7726	-2.7468
0.3812	0.8	100	0.3743	-2.2787	-22.2937	0.4600	20.0150	-263.9586	-64.9924	-2.7884	-2.7491
0.3119	1.2	150	0.3743	-2.2787	-22.2937	0.4600	20.0150	-263.9586	-64.9924	-2.7884	-2.7491
0.3639	1.6	200	0.3743	-2.2787	-22.2937	0.4600	20.0150	-263.9586	-64.9924	-2.7884	-2.7491
0.4332	2.0	250	0.3743	-2.2787	-22.2937	0.4600	20.0150	-263.9586	-64.9924	-2.7884	-2.7491
0.3986	2.4	300	0.3743	-2.2787	-22.2937	0.4600	20.0150	-263.9586	-64.9924	-2.7884	-2.7491
0.3986	2.8	350	0.3743	-2.2787	-22.2937	0.4600	20.0150	-263.9586	-64.9924	-2.7884	-2.7491
0.4505	3.2	400	0.3743	-2.2787	-22.2937	0.4600	20.0150	-263.9586	-64.9924	-2.7884	-2.7491
0.4505	3.6	450	0.3743	-2.2787	-22.2937	0.4600	20.0150	-263.9586	-64.9924	-2.7884	-2.7491
0.4332	4.0	500	0.3743	-2.2787	-22.2937	0.4600	20.0150	-263.9586	-64.9924	-2.7884	-2.7491
0.3292	4.4	550	0.3743	-2.2787	-22.2937	0.4600	20.0150	-263.9586	-64.9924	-2.7884	-2.7491
0.3639	4.8	600	0.3743	-2.2787	-22.2937	0.4600	20.0150	-263.9586	-64.9924	-2.7884	-2.7491
0.4505	5.2	650	0.3743	-2.2787	-22.2937	0.4600	20.0150	-263.9586	-64.9924	-2.7884	-2.7491
0.4505	5.6	700	0.3743	-2.2787	-22.2937	0.4600	20.0150	-263.9586	-64.9924	-2.7884	-2.7491
0.3639	6.0	750	0.3743	-2.2787	-22.2937	0.4600	20.0150	-263.9586	-64.9924	-2.7884	-2.7491
0.2426	6.4	800	0.3743	-2.2787	-22.2937	0.4600	20.0150	-263.9586	-64.9924	-2.7884	-2.7491
0.5025	6.8	850	0.3743	-2.2787	-22.2937	0.4600	20.0150	-263.9586	-64.9924	-2.7884	-2.7491
0.3119	7.2	900	0.3743	-2.2787	-22.2937	0.4600	20.0150	-263.9586	-64.9924	-2.7884	-2.7491
0.3466	7.6	950	0.3743	-2.2787	-22.2937	0.4600	20.0150	-263.9586	-64.9924	-2.7884	-2.7491
0.3812	8.0	1000	0.3743	-2.2787	-22.2937	0.4600	20.0150	-263.9586	-64.9924	-2.7884	-2.7491

Framework versions

Transformers 4.44.2
Pytorch 2.0.0+cu117
Datasets 3.0.0
Tokenizers 0.19.1

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