See axolotl config

axolotl version: 0.4.1

adapter: lora
auto_find_batch_size: false
base_model: migtissera/Tess-v2.5-Phi-3-medium-128k-14B
bf16: auto
chat_template: llama3
dataloader_num_workers: 12
dataset_prepared_path: null
datasets:
- data_files:
  - b476e768184c0288_train_data.json
  ds_type: json
  format: custom
  path: /workspace/input_data/b476e768184c0288_train_data.json
  type:
    field_instruction: src_text
    field_output: tgt_text
    format: '{instruction}'
    no_input_format: '{instruction}'
    system_format: '{system}'
    system_prompt: ''
debug: null
deepspeed: null
early_stopping_patience: 3
early_stopping_threshold: 1.0e-05
eval_max_new_tokens: 128
eval_steps: 141
eval_strategy: null
flash_attention: true
fp16: null
fsdp: null
fsdp_config: null
gradient_accumulation_steps: 2
gradient_checkpointing: false
group_by_length: false
hub_model_id: mrferr3t/b5632a9f-bcb9-4dac-acb1-5049251e1877
hub_repo: null
hub_strategy: checkpoint
hub_token: null
learning_rate: 0.0004
load_in_4bit: false
load_in_8bit: false
local_rank: null
logging_steps: 141
lora_alpha: 16
lora_dropout: 0.05
lora_fan_in_fan_out: null
lora_model_dir: null
lora_r: 8
lora_target_linear: true
lr_scheduler: cosine
max_steps: 
micro_batch_size: 4
mlflow_experiment_name: /tmp/b476e768184c0288_train_data.json
model_type: AutoModelForCausalLM
num_epochs: 100
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: 141
saves_per_epoch: 0
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: null
wandb_mode: 
wandb_name: 37128511-879e-4070-8fa6-483ed98c0cae
wandb_project: Gradients-On-Demand
wandb_run: your_name
wandb_runid: 37128511-879e-4070-8fa6-483ed98c0cae
warmup_steps: 100
weight_decay: 0.0
xformers_attention: null

b5632a9f-bcb9-4dac-acb1-5049251e1877

This model is a fine-tuned version of migtissera/Tess-v2.5-Phi-3-medium-128k-14B on the None dataset. It achieves the following results on the evaluation set:

• Loss: 1.6279

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: 0.0004
• train_batch_size: 4
• eval_batch_size: 4
• seed: 42
• gradient_accumulation_steps: 2
• 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: 100
• num_epochs: 100

Training results

Training Loss	Epoch	Step	Validation Loss
No log	0.0025	1	4.1070
5.5222	0.3579	141	1.9694
3.8869	0.7157	282	1.7862
3.9941	1.0736	423	1.7213
3.0632	1.4315	564	1.6551
3.1662	1.7893	705	1.6314
2.6543	2.1472	846	1.6497
2.0306	2.5051	987	1.6255
2.2417	2.8629	1128	1.5397
1.7098	3.2208	1269	1.6804
1.4168	3.5787	1410	1.7165
1.489	3.9365	1551	1.6279

Framework versions

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