README.md

metadata

base_model: meta-llama/Meta-Llama-3.1-8B-Instruct
datasets:
  - GaetanMichelet/chat-60_ft_task-2
  - GaetanMichelet/chat-120_ft_task-2
  - GaetanMichelet/chat-180_ft_task-2
library_name: peft
license: llama3.1
tags:
  - alignment-handbook
  - trl
  - sft
  - generated_from_trainer
model-index:
  - name: Llama-31-8B_task-2_180-samples_config-3
    results: []

Llama-31-8B_task-2_180-samples_config-3

This model is a fine-tuned version of meta-llama/Meta-Llama-3.1-8B-Instruct on the GaetanMichelet/chat-60_ft_task-2, the GaetanMichelet/chat-120_ft_task-2 and the GaetanMichelet/chat-180_ft_task-2 datasets. It achieves the following results on the evaluation set:

• Loss: 0.7140

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: 42
• distributed_type: multi-GPU
• gradient_accumulation_steps: 8
• total_train_batch_size: 8
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: cosine
• lr_scheduler_warmup_ratio: 0.1
• num_epochs: 150

Training results

Training Loss	Epoch	Step	Validation Loss
1.0365	1.0	17	1.1316
1.1746	2.0	34	1.1196
1.0933	3.0	51	1.0957
0.985	4.0	68	1.0540
0.9741	5.0	85	0.9950
1.0008	6.0	102	0.9377
0.8935	7.0	119	0.8939
0.8862	8.0	136	0.8579
0.8266	9.0	153	0.8294
0.7797	10.0	170	0.8075
0.8158	11.0	187	0.7903
0.6845	12.0	204	0.7742
0.6819	13.0	221	0.7598
0.7241	14.0	238	0.7472
0.695	15.0	255	0.7365
0.6982	16.0	272	0.7272
0.622	17.0	289	0.7215
0.5905	18.0	306	0.7156
0.6121	19.0	323	0.7140
0.567	20.0	340	0.7166
0.5471	21.0	357	0.7172
0.4761	22.0	374	0.7234
0.4967	23.0	391	0.7358
0.4833	24.0	408	0.7644
0.4071	25.0	425	0.8012
0.3567	26.0	442	0.8289

Framework versions

• PEFT 0.12.0
• Transformers 4.44.0
• Pytorch 2.1.2+cu121
• Datasets 2.20.0
• Tokenizers 0.19.1