metadata
license: other
library_name: peft
tags:
- alignment-handbook
- trl
- sft
- generated_from_trainer
- trl
- sft
- generated_from_trainer
datasets:
- nthakur/miracl-raft-sft-instruct-v0.1
- nthakur/nomiracl-raft-sft-instruct-v0.1
- nthakur/miracl-en-x-raft-sft-instruct-v0.1
- nthakur/miracl-x-en-raft-sft-instruct-v0.1
base_model: meta-llama/Meta-Llama-3-8B-Instruct
model-index:
- name: Meta-Llama-3-8B-Instruct-miracl-mix-raft-sft-25th-apr-v1.0
results: []
Meta-Llama-3-8B-Instruct-miracl-mix-raft-sft-25th-apr-v1.0
This model is a fine-tuned version of meta-llama/Meta-Llama-3-8B-Instruct on the nthakur/miracl-raft-sft-instruct-v0.1, the nthakur/nomiracl-raft-sft-instruct-v0.1, the nthakur/miracl-en-x-raft-sft-instruct-v0.1 and the nthakur/miracl-x-en-raft-sft-instruct-v0.1 datasets. It achieves the following results on the evaluation set:
- Loss: 1.3064
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: 2
- seed: 42
- distributed_type: multi-GPU
- num_devices: 4
- gradient_accumulation_steps: 4
- total_train_batch_size: 32
- total_eval_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: 1
Training results
| Training Loss | Epoch | Step | Validation Loss |
|---|---|---|---|
| 1.4903 | 0.09 | 200 | 1.3961 |
| 1.465 | 0.18 | 400 | 1.3499 |
| 1.4193 | 0.28 | 600 | 1.3330 |
| 1.3593 | 0.37 | 800 | 1.3232 |
| 1.3552 | 0.46 | 1000 | 1.3166 |
| 1.3685 | 0.55 | 1200 | 1.3123 |
| 1.3487 | 0.64 | 1400 | 1.3094 |
| 1.3891 | 0.74 | 1600 | 1.3076 |
| 1.3858 | 0.83 | 1800 | 1.3067 |
| 1.3635 | 0.92 | 2000 | 1.3064 |
Framework versions
- PEFT 0.10.0
- Transformers 4.39.0.dev0
- Pytorch 2.1.2+cu121
- Datasets 2.18.0
- Tokenizers 0.15.2