README.md

metadata

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

Llama-31-8B_task-1_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-1, the GaetanMichelet/chat-120_ft_task-1 and the GaetanMichelet/chat-180_ft_task-1 datasets. It achieves the following results on the evaluation set:

• Loss: 1.2578

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
2.1142	1.0	17	2.0665
1.999	2.0	34	2.0373
1.9478	3.0	51	1.9763
1.8469	4.0	68	1.8651
1.7424	5.0	85	1.7083
1.6109	6.0	102	1.5913
1.4629	7.0	119	1.5339
1.4728	8.0	136	1.4790
1.4044	9.0	153	1.4184
1.2885	10.0	170	1.3676
1.3419	11.0	187	1.3272
1.1269	12.0	204	1.3011
1.1586	13.0	221	1.2765
1.1923	14.0	238	1.2690
1.1147	15.0	255	1.2578
1.0264	16.0	272	1.2624
0.9349	17.0	289	1.2617
0.8531	18.0	306	1.2745
0.7922	19.0	323	1.3237
0.7745	20.0	340	1.3550
0.7142	21.0	357	1.4246
0.4873	22.0	374	1.5460

Framework versions

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