See axolotl config

axolotl version: 0.6.0

# Original base model config
# base_model: Dans-DiscountModels/Meta-Llama-3.2-3B-ChatML
# Using smaller model instead
base_model: Emm9625/Llama-3.2-1B-chatml

# Original tokenizer config
# tokenizer_config: Dans-DiscountModels/Meta-Llama-3.2-3B-ChatML
# Using matching tokenizer for smaller model
tokenizer_config: Emm9625/Llama-3.2-1B-chatml

# Model loading configuration
load_in_8bit: false
load_in_4bit: false
strict: false

# Chat template configuration
chat_template: chatml

# Dataset configuration
datasets:
  - path: Emm9625/textwork-00
    name: smol-constraints
    split: train
    type: chat_template
    field_messages: messages
    message_field_role: role
    message_field_content: content
    train_on_eos: turn
    # shards: 2
    # shard_idx: 0
    
  - path: Emm9625/textwork-00
    name: smol-rewrite
    split: train
    type: chat_template
    field_messages: messages
    message_field_role: role
    message_field_content: content
    train_on_eos: turn
    # shards: 2
    # shard_idx: 0
    
  - path: Emm9625/textwork-00
    name: smol-summarize
    split: train
    type: chat_template
    field_messages: messages
    message_field_role: role
    message_field_content: content
    train_on_eos: turn
    # shards: 2
    # shard_idx: 0
    

test_datasets:
  - path: Emm9625/textwork-00
    name: smol-constraints
    split: test
    type: chat_template
    field_messages: messages
    message_field_role: role
    message_field_content: content
    train_on_eos: turn
    shards: 10
    shard_idx: 0
    
  - path: Emm9625/textwork-00
    name: smol-rewrite
    split: test
    type: chat_template
    field_messages: messages
    message_field_role: role
    message_field_content: content
    train_on_eos: turn
    shards: 10
    shard_idx: 0

  - path: Emm9625/textwork-00
    name: smol-summarize
    split: test
    type: chat_template
    field_messages: messages
    message_field_role: role
    message_field_content: content
    train_on_eos: turn
    shards: 10
    shard_idx: 0



dataset_prepared_path: last_run_prepared
output_dir: /tmp/meow/
hub_model_id: Emm9625/textwork-00-1B-25-01-18
hub_strategy: checkpoint
# Whether to use hf `use_auth_token` for loading datasets. Useful for fetching private datasets
# Required to be true when used in combination with `push_dataset_to_hub`
hf_use_auth_token: true

# Model configuration
sequence_len: 4096
sample_packing: true
eval_sample_packing: true
pad_to_sequence_len: true

adapter:
lora_model_dir:

lora_r: 16
lora_alpha: 32
lora_dropout: 0.05
lora_fan_in_fan_out:
lora_target_modules:
  - gate_proj
  - down_proj
  - up_proj
  - q_proj
  - v_proj
  - k_proj
  - o_proj


# Unsloth optimizations
unsloth_cross_entropy_loss: true
unsloth_rms_norm: true
unsloth_rope: true
#Lora Optimizations
# unsloth_lora_mlp: true
# unsloth_lora_qkv: true
# unsloth_lora_o: true



# Training configuration
gradient_accumulation_steps: 2
micro_batch_size: 8
num_epochs: 1
optimizer: adamw_8bit
lr_scheduler: cosine
learning_rate: 2e-5
torch_compile: auto

train_on_inputs: false
group_by_length: false
bf16: true
gradient_checkpointing: true
flash_attention: true

# Training monitoring
loss_watchdog_threshold: 5.0
loss_watchdog_patience: 3
warmup_ratio: 0.1
weight_decay: 0.00
saves_per_epoch: 1
evals_per_epoch: 5
save_safetensors: true
wandb_project: textwork-01
logging_steps: 1

# Special tokens configuration
special_tokens:
  eos_token: "<|im_end|>"
  bos_token: "<|im_start|>"

fsdp:
fsdp_config:

textwork-00-1B-25-01-18

This model is a fine-tuned version of Emm9625/Llama-3.2-1B-chatml on the Emm9625/textwork-00, the Emm9625/textwork-00 and the Emm9625/textwork-00 datasets. It achieves the following results on the evaluation set:

• Loss: 0.6892

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: 2e-05
• train_batch_size: 8
• eval_batch_size: 8
• seed: 42
• gradient_accumulation_steps: 2
• total_train_batch_size: 16
• optimizer: Use OptimizerNames.ADAMW_8BIT 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: 109
• num_epochs: 1

Training results

Training Loss	Epoch	Step	Validation Loss
1.3666	0.0009	1	1.3578
0.7961	0.2003	220	0.7602
0.7336	0.4005	440	0.7124
0.7415	0.6008	660	0.6946
0.7035	0.8011	880	0.6892

Framework versions

• Transformers 4.47.1
• Pytorch 2.5.1+cu124
• Datasets 3.2.0
• Tokenizers 0.21.0