Built with Axolotl

See axolotl config

axolotl version: 0.5.2

adapter: lora
base_model: unsloth/mistral-7b
bf16: auto
datasets:
- data_files:
  - 8219297a1f15c78f_train_data.json
  ds_type: json
  format: custom
  path: 8219297a1f15c78f_train_data.json
  preprocessing:
  - shuffle: true
  type:
    field: null
    field_input: null
    field_instruction: prompt
    field_output: response
    field_system: null
    format: null
    no_input_format: null
    system_format: '{system}'
    system_prompt: ''
debug: null
device_map: auto
early_stopping_patience: null
eval_max_new_tokens: 16
eval_table_max_new_tokens: 128
eval_table_size: null
evals_per_epoch: 8
flash_attention: true
fp16: null
fsdp: null
fsdp_config: null
gradient_accumulation_steps: 4
gradient_checkpointing: true
group_by_length: false
hub_model_id: taopanda/test-mistral-7b
is_mistral_derived_model: true
learning_rate: 0.0005
load_in_4bit: false
load_in_8bit: false
local_rank: null
logging_steps: 1
lora_alpha: 128
lora_dropout: 0.05
lora_r: 64
lora_target_modules:
- q_proj
- v_proj
lr_scheduler: cosine
max_grad_norm: 1.0
max_steps: 450
micro_batch_size: 4
model_type: MistralForCausalLM
num_epochs: 1
optimizer: adamw_torch
output_dir: ./out/taopanda_test-mistral-7b
pad_to_sequence_len: true
resume_from_checkpoint: null
save_steps: 0.15
save_total_limit: 1
seed: 42
sequence_len: 1024
special_tokens:
  bos_token: <s>
  eos_token: </s>
  unk_token: <unk>
strict: false
tf32: true
tokenizer_type: AutoTokenizer
train_on_inputs: false
trust_remote_code: true
use_flash_attn_2: true
val_set_size: 0.1
wandb_entity: fatcat87-taopanda
wandb_log_model: null
wandb_mode: online
wandb_name: taopanda_test-mistral-7b
wandb_project: subnet56-test
wandb_runid: taopanda_test-mistral-7b
wandb_watch: null
warmup_ratio: 0.06
weight_decay: 0.01
xformers_attention: null

test-mistral-7b

This model is a fine-tuned version of unsloth/mistral-7b on the None dataset. It achieves the following results on the evaluation set:

  • Loss: 0.6985

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: 0.0005
  • train_batch_size: 4
  • eval_batch_size: 4
  • seed: 42
  • distributed_type: multi-GPU
  • num_devices: 4
  • gradient_accumulation_steps: 4
  • total_train_batch_size: 64
  • total_eval_batch_size: 16
  • optimizer: Use OptimizerNames.ADAMW_TORCH with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
  • lr_scheduler_type: cosine
  • lr_scheduler_warmup_steps: 5
  • training_steps: 87

Training results

Training Loss Epoch Step Validation Loss
1.0651 0.0116 1 1.0560
0.7617 0.1272 11 0.7764
0.7784 0.2543 22 0.7421
0.6848 0.3815 33 0.7290
0.7263 0.5087 44 0.7161
0.7062 0.6358 55 0.7074
0.7281 0.7630 66 0.7020
0.7402 0.8902 77 0.6985

Framework versions

  • PEFT 0.13.2
  • Transformers 4.46.3
  • Pytorch 2.5.1+cu124
  • Datasets 3.1.0
  • Tokenizers 0.20.3
Downloads last month
1
Inference Providers NEW
This model isn't deployed by any Inference Provider. ๐Ÿ™‹ Ask for provider support

Model tree for taopanda/test-mistral-7b

Base model

unsloth/mistral-7b
Adapter
(218)
this model