See axolotl config

axolotl version: 0.4.1

adapter: lora
base_model: unsloth/gemma-2-2b-it
bf16: true
chat_template: llama3
dataset_prepared_path: null
datasets:
- data_files:
  - 5dc5562c268ed6f8_train_data.json
  ds_type: json
  format: custom
  path: /workspace/input_data/5dc5562c268ed6f8_train_data.json
  type:
    field_input: original_version
    field_instruction: title
    field_output: french_version
    format: '{instruction} {input}'
    no_input_format: '{instruction}'
    system_format: '{system}'
    system_prompt: ''
debug: null
device_map:
  ? ''
  : 0,1,2,3,4,5,6,7
early_stopping_patience: 2
eval_max_new_tokens: 128
eval_steps: 400
eval_table_size: null
flash_attention: false
gradient_accumulation_steps: 4
gradient_checkpointing: true
group_by_length: false
hub_model_id: Alphatao/f11e9dd7-c7de-4f97-9479-ba5c0483a0e4
hub_repo: null
hub_strategy: null
hub_token: null
learning_rate: 0.0002
load_best_model_at_end: true
load_in_4bit: false
load_in_8bit: false
local_rank: null
logging_steps: 1
lora_alpha: 32
lora_dropout: 0.05
lora_fan_in_fan_out: null
lora_model_dir: null
lora_r: 16
lora_target_linear: true
lora_target_modules:
- q_proj
- k_proj
- v_proj
- o_proj
lr_scheduler: cosine
max_grad_norm: 1.0
max_steps: 7258
micro_batch_size: 2
mlflow_experiment_name: /tmp/5dc5562c268ed6f8_train_data.json
model_type: AutoModelForCausalLM
num_epochs: 2
optimizer: adamw_bnb_8bit
output_dir: miner_id_24
pad_to_sequence_len: true
resume_from_checkpoint: null
s2_attention: null
sample_packing: false
save_steps: 400
sequence_len: 1024
strict: false
tf32: true
tokenizer_type: AutoTokenizer
train_on_inputs: false
trust_remote_code: true
val_set_size: 0.05
wandb_entity: null
wandb_mode: online
wandb_name: 4ed3c119-858f-4aa1-a7e8-0ec4deb92bdb
wandb_project: Gradients-On-Demand
wandb_run: your_name
wandb_runid: 4ed3c119-858f-4aa1-a7e8-0ec4deb92bdb
warmup_steps: 10
weight_decay: 0.0
xformers_attention: null

f11e9dd7-c7de-4f97-9479-ba5c0483a0e4

This model is a fine-tuned version of unsloth/gemma-2-2b-it on the None dataset. It achieves the following results on the evaluation set:

• Loss: 0.9316

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
• gradient_accumulation_steps: 4
• total_train_batch_size: 8
• 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: 7258

Training results

Training Loss	Epoch	Step	Validation Loss
0.9599	0.0001	1	1.3887
1.1727	0.0342	400	1.0594
0.4701	0.0684	800	1.0372
0.909	0.1026	1200	1.0207
1.2693	0.1368	1600	1.0106
1.0021	0.1710	2000	1.0011
0.6727	0.2052	2400	0.9943
0.9811	0.2394	2800	0.9838
1.3031	0.2736	3200	0.9744
0.5804	0.3078	3600	0.9673
0.6857	0.3420	4000	0.9602
0.8827	0.3762	4400	0.9520
1.0814	0.4104	4800	0.9460
0.4212	0.4446	5200	0.9410
1.0149	0.4788	5600	0.9374
0.6131	0.5130	6000	0.9345
0.6607	0.5472	6400	0.9326
0.8703	0.5814	6800	0.9318
1.2847	0.6156	7200	0.9316

Framework versions

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