Guten-SFT-LORA

<!-- This model card has been generated automatically according to the information the Trainer had access to. You should probably proofread and complete it, then remove this comment. -->

<img src="https://raw.githubusercontent.com/axolotl-ai-cloud/axolotl/main/image/axolotl-badge-web.png" alt="Built with Axolotl" width="200" height="32"/>

<details><summary>See axolotl config</summary>

axolotl version: 0.13.2

base_model: allenai/Olmo-3.1-32B-Instruct

plugins:
  - axolotl.integrations.liger.LigerPlugin

liger_rope: true
liger_rms_norm: true
liger_glu_activation: true
liger_layer_norm: true
liger_fused_linear_cross_entropy: true

lora_mlp_kernel: false
lora_qkv_kernel: false
lora_o_kernel: false

load_in_8bit: true
load_in_4bit: false

sequence_len: 6144
max_sample_length: 6144

sample_packing: true
gradient_checkpointing: true
flash_attention: true


bf16: true
tf32: true


datasets:
  - path: ./gutenbeg-2-sft-train.jsonl
    type: chat_template
    train_on_eos: 'all'

  - path: ./condor-sft-train.jsonl
    type: chat_template
    train_on_eos: 'all'

chat_template: chatml
eot_tokens: 
    - "<|im_end|>"

adapter: lora
lora_r: 64
lora_alpha: 128
lora_dropout: 0.0
lora_bias: None
lora_target_linear: true
use_tensorboard: true

optimizer: paged_adamw_8bit
learning_rate: 2.5e-5
loraplus_lr_ratio: 16

# Training arguments
output_dir: ./Olmo-Guten-SFT
num_epochs: 2
micro_batch_size: 2
gradient_accumulation_steps: 4
warmup_ratio: 0.05
lr_scheduler: 'constant_with_warmup'
max_grad_norm: 1
logging_steps: 1
seed: 42

</details><br>

Olmo-Guten-SFT

This model is a fine-tuned version of allenai/Olmo-3.1-32B-Instruct on the ./gutenbeg-2-sft-train.jsonl and the ./condor-sft-train.jsonl 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: 2.5e-05
• train_batch_size: 2
• eval_batch_size: 2
• seed: 42
• gradient_accumulation_steps: 4
• total_train_batch_size: 8
• optimizer: Use OptimizerNames.PAGED_ADAMW_8BIT with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: constant_with_warmup
• lr_scheduler_warmup_steps: 2
• training_steps: 42

Training results

Framework versions

• PEFT 0.18.1
• Transformers 4.57.6
• Pytorch 2.8.0+cu128
• Datasets 4.5.0
• Tokenizers 0.22.2