Built with Axolotl

See axolotl config

axolotl version: 0.4.0

base_model: T3Q-LLM/T3Q-LLM-sft1.0-dpo1.0
base_model_config: T3Q-LLM/T3Q-LLM-sft1.0-dpo1.0
model_type: AutoModelForCausalLM
tokenizer_type: AutoTokenizer
is_llama_derived_model: true
hub_model_id: T3Q-LLM-sft1.0-dpo1.0_100QA

load_in_8bit: false
load_in_4bit: true
strict: false

datasets:
  - path: admin_data.csv
    type: alpaca
      # The below are defaults. only set what's needed if you use a different column name.
      # system_prompt: ""
      # system_format: "{system}"
      # field_system: system
      # field_instruction: instruction
      # field_input: input
      # field_output: output

      # format: |-
      #   Human: {instruction} {input}
      #   Assistant:

      # no_input_format: "{instruction} "

# dataset_prepared_path: yanolja_preprocessed_data
dataset_prepared_path: last_run_prepared
val_set_size: 0.2
output_dir: ./T3Q-LLM-sft1.0-dpo1.0_100QA

adapter: qlora
lora_model_dir: 

# device_map: [0,1,3]

sequence_len: 4096
sample_packing: false

lora_r: 32
lora_alpha: 16
lora_dropout: 0.05
lora_target_modules:
lora_target_linear: true
lora_fan_in_fan_out: 

wandb_project: axolotl_T3Q
wandb_entity: 
wandb_watch: 
wandb_run_id: T3Q_mod_100
wandb_log_model: 

gradient_accumulation_steps: 4
micro_batch_size: 2
num_epochs: 10
optimizer: paged_adamw_32bit
lr_scheduler: cosine
learning_rate: 0.0002

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

gradient_checkpointing: true
early_stopping_patience: 
resume_from_checkpoint: 
local_rank: 
logging_steps: 1
xformers_attention: 
flash_attention: true

warmup_steps: 100
eval_steps: 0.01
save_strategy: epoch
save_steps: 
debug: 
deepspeed: 
weight_decay: 0.0
fsdp: 
fsdp_config: 
special_tokens:
    bos_token: "<s>"
    eos_token: "<|im_end|>"
    unk_token: "<unk>"
    pad_token: "</s>"  # EOS와 PAD가 동일

T3Q-LLM-sft1.0-dpo1.0_100QA

This model is a fine-tuned version of T3Q-LLM/T3Q-LLM-sft1.0-dpo1.0 on the None dataset. It achieves the following results on the evaluation set:

  • Loss: 0.6629

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: 2
  • eval_batch_size: 2
  • seed: 42
  • distributed_type: multi-GPU
  • num_devices: 2
  • gradient_accumulation_steps: 4
  • total_train_batch_size: 16
  • total_eval_batch_size: 4
  • optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
  • lr_scheduler_type: cosine
  • lr_scheduler_warmup_steps: 100
  • num_epochs: 10

Training results

Training Loss Epoch Step Validation Loss
0.8867 0.2 1 1.0564
0.9385 0.4 2 1.0557
0.9454 0.6 3 1.0535
0.8469 0.8 4 1.0494
0.8583 1.0 5 1.0412
0.8691 1.2 6 1.0262
0.8306 1.4 7 1.0073
0.8302 1.6 8 0.9834
0.8028 1.8 9 0.9556
0.8987 2.0 10 0.9181
0.7826 2.2 11 0.8777
0.6936 2.4 12 0.8379
0.6453 2.6 13 0.8035
0.6613 2.8 14 0.7741
0.6548 3.0 15 0.7483
0.6078 3.2 16 0.7238
0.6185 3.4 17 0.7004
0.5293 3.6 18 0.6815
0.5617 3.8 19 0.6666
0.4845 4.0 20 0.6541
0.4904 4.2 21 0.6443
0.5375 4.4 22 0.6349
0.5099 4.6 23 0.6254
0.4286 4.8 24 0.6187
0.4952 5.0 25 0.6133
0.4394 5.2 26 0.6089
0.4974 5.4 27 0.6041
0.3877 5.6 28 0.5999
0.4992 5.8 29 0.5952
0.4187 6.0 30 0.5902
0.4302 6.2 31 0.5871
0.3861 6.4 32 0.5836
0.3966 6.6 33 0.5805
0.4399 6.8 34 0.5786
0.3732 7.0 35 0.5777
0.3727 7.2 36 0.5780
0.3442 7.4 37 0.5786
0.3477 7.6 38 0.5801
0.3763 7.8 39 0.5808
0.3498 8.0 40 0.5824
0.312 8.2 41 0.5834
0.3282 8.4 42 0.5869
0.2938 8.6 43 0.5912
0.2908 8.8 44 0.5967
0.3083 9.0 45 0.6031
0.244 9.2 46 0.6111
0.2894 9.4 47 0.6228
0.2318 9.6 48 0.6353
0.2375 9.8 49 0.6474
0.1939 10.0 50 0.6629

Framework versions

  • PEFT 0.10.0
  • Transformers 4.40.1
  • Pytorch 2.1.2+cu121
  • Datasets 2.15.0
  • Tokenizers 0.19.1
Downloads last month
2
Safetensors
Model size
10.8B params
Tensor type
FP16
·
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 superiort/T3Q-LLM-sft1.0-dpo1.0_100QA_10epochs

Adapter
(2)
this model