YoungPanda/reg_eu_llama3

See axolotl config

axolotl version: 0.4.0

base_model: winglian/Llama-3-8b-64k-PoSE
model_type: LlamaForCausalLM
tokenizer_type: AutoTokenizer

load_in_8bit: false
load_in_4bit: false
strict: false

datasets:
  - path: eu_regulatory_ir
    name: eu2uk
    type: completion
dataset_prepared_path: last_run_prepared
val_set_size: 0.05
output_dir: ./out

sequence_len: 4096
sample_packing: true
pad_to_sequence_len: true
eval_sample_packing: false
wandb_project: eullama
wandb_entity:
wandb_watch:
wandb_name:
wandb_log_model:

gradient_accumulation_steps: 1
micro_batch_size: 4
num_epochs: 3
optimizer: paged_adamw_8bit
lr_scheduler: cosine
learning_rate: 2e-5

train_on_inputs: false
group_by_length: false
bf16: auto
fp16:
tf32: false

gradient_checkpointing: true
gradient_checkpointing_kwargs:
  use_reentrant: false
early_stopping_patience:
resume_from_checkpoint:
logging_steps: 1
xformers_attention:
flash_attention: true

warmup_steps: 100
evals_per_epoch: 2
eval_table_size:
saves_per_epoch: 1
debug:
deepspeed:
weight_decay: 0.0
fsdp:
fsdp_config:
special_tokens:
  pad_token: <|end_of_text|>

out

This model is a fine-tuned version of winglian/Llama-3-8b-64k-PoSE on the None dataset. It achieves the following results on the evaluation set:

• Loss: 1.0796

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: 2e-05
• train_batch_size: 4
• eval_batch_size: 4
• seed: 42
• distributed_type: multi-GPU
• num_devices: 7
• total_train_batch_size: 28
• total_eval_batch_size: 28
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: cosine
• lr_scheduler_warmup_steps: 100
• num_epochs: 3

Training results

Training Loss	Epoch	Step	Validation Loss
1.2827	0.03	1	1.2008
1.2646	0.51	18	1.1367
1.1913	1.03	36	1.0856
1.0415	1.49	54	1.0675
0.9863	2.0	72	1.0485
0.7384	2.46	90	1.0796

Framework versions

• Transformers 4.40.0.dev0
• Pytorch 2.2.0+cu121
• Datasets 2.15.0
• Tokenizers 0.15.0