Built with Axolotl

See axolotl config

axolotl version: 0.4.1

base_model: SeaLLMs/SeaLLM3-7B-Chat
trust_remote_code: true

load_in_8bit: false
load_in_4bit: true
strict: false

datasets:
  - path: Tippawan/p9-seallm
    type: sharegpt
    conversation: chatml
    field_messages: messages
chat_template: chatml
dataset_prepared_path:
val_set_size: 0.00 #editted 2
output_dir: ./outputs/outputs_name

sequence_len: 2048
sample_packing: false
eval_sample_packing: false
pad_to_sequence_len: false

push_to_hub: true
hub_model_id: Tippawan/proof-reading-SeaLLM3-7B-Chat-3090-v9  # Replace with your Hugging Face repo ID
use_auth_token: true  # Ensure you have set your Hugging Face API token in the environment
hub_private_repo: true  # Set to true if you want the repository to be private
hub_strategy: all_checkpoints
save_total_limit: 3
load_best_model_at_end: true

adapter: lora
lora_model_dir: Tippawan/proof-reading-SeaLLM3-7B-Chat-3090-v8
lora_r: 16
lora_alpha: 32
lora_dropout: 0.05
lora_target_linear: true
lora_fan_in_fan_out:

wandb_project: proof-reading-SeaLLM3-7B-Chat-3090-v9
wandb_entity: 
wandb_watch:
wandb_name:
wandb_log_model:

gradient_accumulation_steps: 4
micro_batch_size: 8
num_epochs: 3 #editted 3
optimizer: adamw_torch
lr_scheduler: cosine
learning_rate: 0.0002

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

gradient_checkpointing: true
early_stopping_patience:
resume_from_checkpoint:
local_rank:
logging_steps: 1
xformers_attention:
flash_attention: true
seed: 42
warmup_steps: 10
evals_per_epoch: 1
saves_per_epoch: 1
debug:
deepspeed:
weight_decay: 0.0
fsdp:
fsdp_config:
special_tokens:

proof-reading-SeaLLM3-7B-Chat-3090-v9

This model is a fine-tuned version of SeaLLMs/SeaLLM3-7B-Chat on the None dataset.

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: 0.0002
  • train_batch_size: 8
  • eval_batch_size: 8
  • seed: 42
  • gradient_accumulation_steps: 4
  • total_train_batch_size: 32
  • optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
  • lr_scheduler_type: cosine
  • lr_scheduler_warmup_steps: 10
  • num_epochs: 3

Training results

Framework versions

  • PEFT 0.12.0
  • Transformers 4.45.0.dev0
  • Pytorch 2.3.1+cu121
  • Datasets 2.21.0
  • Tokenizers 0.19.1
Downloads last month
14
Inference Providers NEW
This model is not currently available via any of the supported Inference Providers.
The model cannot be deployed to the HF Inference API: The model has no pipeline_tag.

Model tree for Tippawan/proof-reading-SeaLLM3-7B-Chat-3090-v9

Adapter
(10)
this model