See axolotl config

axolotl version: 0.4.1

adapter: lora
base_model: katuni4ka/tiny-random-qwen1.5-moe
bf16: auto
chat_template: llama3
dataset_prepared_path: null
datasets:
- data_files:
  - 22c6d07080dfffd0_train_data.json
  ds_type: json
  format: custom
  path: /workspace/input_data/22c6d07080dfffd0_train_data.json
  type:
    field_instruction: text
    field_output: label_text
    format: '{instruction}'
    no_input_format: '{instruction}'
    system_format: '{system}'
    system_prompt: ''
debug: null
deepspeed: null
early_stopping_patience: null
eval_max_new_tokens: 128
eval_table_size: null
evals_per_epoch: 4
flash_attention: false
fp16: null
fsdp: null
fsdp_config: null
gradient_accumulation_steps: 8
gradient_checkpointing: false
group_by_length: false
hub_model_id: leixa/f46cc6fc-f320-4e60-88e9-6f617fd7b16a
hub_repo: null
hub_strategy: checkpoint
hub_token: null
learning_rate: 0.0001
load_in_4bit: false
load_in_8bit: false
local_rank: 0
logging_steps: 3
lora_alpha: 128
lora_dropout: 0.1
lora_fan_in_fan_out: true
lora_model_dir: null
lora_r: 64
lora_target_linear: true
lr_scheduler: cosine
max_steps: 500
micro_batch_size: 4
mlflow_experiment_name: /tmp/22c6d07080dfffd0_train_data.json
model_type: AutoModelForCausalLM
num_epochs: 3
optimizer: adamw_bnb_8bit
output_dir: miner_id_24
pad_to_sequence_len: true
resume_from_checkpoint: null
s2_attention: false
sample_packing: false
saves_per_epoch: 4
sequence_len: 512
strict: false
tf32: false
tokenizer_type: AutoTokenizer
train_on_inputs: false
trust_remote_code: true
val_set_size: 0.05
wandb_entity: leixa-personal
wandb_mode: online
wandb_name: f46cc6fc-f320-4e60-88e9-6f617fd7b16a
wandb_project: Gradients-On-Demand
wandb_run: your_name
wandb_runid: f46cc6fc-f320-4e60-88e9-6f617fd7b16a
warmup_steps: 10
weight_decay: 0.01
xformers_attention: null

f46cc6fc-f320-4e60-88e9-6f617fd7b16a

This model is a fine-tuned version of katuni4ka/tiny-random-qwen1.5-moe on the None dataset. It achieves the following results on the evaluation set:

• Loss: 11.4434

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: 0.0001
• train_batch_size: 4
• eval_batch_size: 4
• seed: 42
• gradient_accumulation_steps: 8
• total_train_batch_size: 32
• optimizer: Use OptimizerNames.ADAMW_BNB 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: 10
• training_steps: 500

Training results

Training Loss	Epoch	Step	Validation Loss
No log	0.0005	1	11.9442
11.642	0.0189	42	11.6242
11.4712	0.0378	84	11.4749
11.4696	0.0567	126	11.4634
11.4574	0.0757	168	11.4591
11.4622	0.0946	210	11.4558
11.4469	0.1135	252	11.4504
11.4435	0.1324	294	11.4483
11.4324	0.1513	336	11.4461
11.4376	0.1702	378	11.4444
11.441	0.1891	420	11.4436
11.4425	0.2080	462	11.4434

Framework versions

• PEFT 0.13.2
• Transformers 4.46.0
• Pytorch 2.5.0+cu124
• Datasets 3.0.1
• Tokenizers 0.20.1