See axolotl config

axolotl version: 0.4.1

adapter: lora
auto_find_batch_size: true
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:
  - 6588e3dccd54f9a1_train_data.json
  ds_type: json
  format: custom
  path: /workspace/input_data/6588e3dccd54f9a1_train_data.json
  type:
    field_input: text
    field_instruction: prompt
    field_output: completion
    format: '{instruction} {input}'
    no_input_format: '{instruction}'
    system_format: '{system}'
    system_prompt: ''
debug: null
deepspeed: null
early_stopping_patience: 3
early_stopping_threshold: 0.0001
eval_max_new_tokens: 128
eval_steps: 80
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/c19cd208-4359-4692-a0f4-50bac6a6f881
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: 80
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: 32
mlflow_experiment_name: /tmp/6588e3dccd54f9a1_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: 
s2_attention: null
sample_packing: false
save_steps: 80
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: 46a4a112-eb3e-4209-864d-e697f32e697e
wandb_project: Gradients-On-Demand
wandb_run: your_name
wandb_runid: 46a4a112-eb3e-4209-864d-e697f32e697e
warmup_steps: 100
weight_decay: 0.0
xformers_attention: null

c19cd208-4359-4692-a0f4-50bac6a6f881

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: 0.4056

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: 32
• eval_batch_size: 32
• seed: 42
• gradient_accumulation_steps: 2
• total_train_batch_size: 64
• 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	0.0008	1	1.0582
1.391	0.0654	80	0.4968
0.9365	0.1309	160	0.4548
0.8891	0.1963	240	0.4451
0.8578	0.2618	320	0.4365
0.8491	0.3272	400	0.4250
0.8306	0.3926	480	0.4243
0.8172	0.4581	560	0.4164
0.815	0.5235	640	0.4102
0.7949	0.5890	720	0.4095
0.7857	0.6544	800	0.4036
0.8036	0.7198	880	0.4050
0.7713	0.7853	960	0.3984
0.7684	0.8507	1040	0.3995
0.7746	0.9162	1120	0.3939
0.7654	0.9816	1200	0.3918
0.6469	1.0470	1280	0.4066
0.6177	1.1125	1360	0.4067
0.6122	1.1779	1440	0.4056

Framework versions

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