See axolotl config

axolotl version: 0.12.0.dev0

base_model: model

hub_model_id: hardlyworking/MS32-3
hub_strategy: "all_checkpoints"
push_dataset_to_hub:
hf_use_auth_token: true

plugins:
  - axolotl.integrations.liger.LigerPlugin
  - axolotl.integrations.cut_cross_entropy.CutCrossEntropyPlugin
liger_rope: true
liger_rms_norm: true
liger_layer_norm: true
liger_glu_activation: true
liger_fused_linear_cross_entropy: false
cut_cross_entropy: true

load_in_8bit: false
load_in_4bit: true

chat_template: mistral_v7_tekken
datasets:
  - path: PocketDoc/Dans-Prosemaxx-RepRemover-1
    type: chat_template
    split: train
    field_messages: conversations
    message_property_mappings:
      role: from
      content: value
      user: human
      assistant: gpt
  - path: PocketDoc/Dans-Prosemaxx-InstructWriter-ZeroShot-2
    type: chat_template
    split: train
    field_messages: conversations
    message_property_mappings:
      role: from
      content: value
      user: human
      assistant: gpt
  - path: Delta-Vector/Orion-Personamaxx-RP-Deslopped-V1
    type: chat_template
    split: train
    field_messages: conversations
    message_property_mappings:
      role: from
      content: value
      user: human
      assistant: gpt
  - path: Delta-Vector/Orion-PIPPA-Cleaned-V2
    type: chat_template
    split: train
    field_messages: conversations
    message_property_mappings:
      role: from
      content: value
      user: human
      assistant: gpt
  - path: Delta-Vector/Orion-Creative_Writing-Complexity
    type: chat_template
    split: train
    field_messages: conversations
    message_property_mappings:
      role: from
      content: value
      user: human
      assistant: gpt
  - path: Delta-Vector/Orion-OpenCAI-ShareGPT
    type: chat_template
    split: train
    field_messages: conversations
    message_property_mappings:
      role: from
      content: value
      user: human
      assistant: gpt
  - path: GreenerPastures/AllYourBase5k
    type: chat_template
    split: train
    field_messages: conversations
    message_property_mappings:
      role: from
      content: value
      user: human
      assistant: gpt

val_set_size: 0.0
output_dir: ./outputs/out

adapter: qlora
lora_r: 64
lora_alpha: 32
lora_dropout: 0.0
lora_target_linear: true
peft_use_rslora: true


sequence_len: 8192
sample_packing: true
eval_sample_packing: true
pad_to_sequence_len: true

wandb_project: MS32-3
wandb_entity:
wandb_watch:
wandb_name: MS32-3
wandb_log_model:

gradient_accumulation_steps: 32
micro_batch_size: 1
num_epochs: 1
optimizer: adamw_bnb_8bit
lr_scheduler: cosine
learning_rate: 4e-5
max_grad_norm: 1.0

bf16: auto
tf32: true

gradient_checkpointing: true
gradient_checkpointing_kwargs:
  use_reentrant: false
  unsloth: true
resume_from_checkpoint:
logging_steps: 1
flash_attention: true

warmup_ratio: 0.1
evals_per_epoch:
saves_per_epoch: 4
weight_decay: 0.0025
special_tokens:

MS32-3

This model was trained from scratch on the PocketDoc/Dans-Prosemaxx-RepRemover-1, the PocketDoc/Dans-Prosemaxx-InstructWriter-ZeroShot-2, the Delta-Vector/Orion-Personamaxx-RP-Deslopped-V1, the Delta-Vector/Orion-PIPPA-Cleaned-V2, the Delta-Vector/Orion-Creative_Writing-Complexity, the Delta-Vector/Orion-OpenCAI-ShareGPT and the GreenerPastures/AllYourBase5k datasets.

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: 4e-05
• train_batch_size: 1
• eval_batch_size: 1
• seed: 42
• gradient_accumulation_steps: 32
• total_train_batch_size: 32
• 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: 100

Training results

Framework versions

• PEFT 0.17.0
• Transformers 4.55.0
• Pytorch 2.7.1+cu126
• Datasets 4.0.0
• Tokenizers 0.21.4