See axolotl config

axolotl version: 0.4.1

adapter: lora
auto_find_batch_size: false
base_model: NousResearch/Nous-Hermes-2-SOLAR-10.7B
bf16: auto
chat_template: llama3
dataloader_num_workers: 12
dataset_prepared_path: null
datasets:
- data_files:
  - f580edf688d697f0_train_data.json
  ds_type: json
  format: custom
  path: /workspace/input_data/f580edf688d697f0_train_data.json
  type:
    field_input: essay
    field_instruction: prompt
    field_output: evaluation
    format: '{instruction} {input}'
    no_input_format: '{instruction}'
    system_format: '{system}'
    system_prompt: ''
debug: null
deepspeed: null
early_stopping_patience: 2
early_stopping_threshold: 1.0e-05
eval_max_new_tokens: 128
eval_steps: 246
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/d0f5e564-cc3b-4034-ab31-f200eada7170
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: 246
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: null
micro_batch_size: 4
mlflow_experiment_name: /tmp/f580edf688d697f0_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: /workspace/hub_repo/last-checkpoint
s2_attention: null
sample_packing: false
save_steps: 246
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: .05000000
wandb_entity: null
wandb_mode: disabled
wandb_name: 25202894-ba9f-4341-936a-2f5d69bcd96c
wandb_project: Gradients-On-Demand
wandb_run: your_name
wandb_runid: 25202894-ba9f-4341-936a-2f5d69bcd96c
warmup_steps: 100
weight_decay: 0.0
xformers_attention: null

d0f5e564-cc3b-4034-ab31-f200eada7170

This model is a fine-tuned version of NousResearch/Nous-Hermes-2-SOLAR-10.7B on the None dataset. It achieves the following results on the evaluation set:

• Loss: 0.3959

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	0.0008	1	0.9526
0.4473	0.1996	240	0.3984
0.3952	0.3992	480	0.3885
0.391	0.5988	720	0.3881
0.3919	0.7983	960	0.3836
0.3881	0.9979	1200	0.3851
0.3575	1.1975	1440	0.3911
0.3677	1.3975	1680	0.3909
0.3832	1.5971	1920	0.3959

Framework versions

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