eac123/oocr-exp1b-e2

See axolotl config

axolotl version: 0.15.0

# ── Continued Pretraining: 7B on 8×A40 (48GB) ──

base_model: allenai/Olmo-3-1025-7B
tokenizer_type: AutoTokenizer

# ── Data ──
datasets:
  - path: data/1b/all.jsonl
    type: completion
    field: completion

# ── Sequence / packing ──
sequence_len: 2048
sample_packing: true
pad_to_sequence_len: true
# NOTE: do NOT enable group_by_length with sample_packing

# ── Batch sizing ──
# Per-GPU: 4 seqs × 2048 tok = 8k tokens/step/GPU
# Global:  4 × 4 accum × 8 GPUs = 128 effective seqs/step
micro_batch_size: 4
gradient_accumulation_steps: 4

# ── Training ──
train_on_inputs: true
optimizer: adamw_torch
lr_scheduler: cosine
learning_rate: 5e-5
warmup_steps: 200
max_steps: 150
weight_decay: 0.01

# ── Precision / memory ──
bf16: true
flash_attention: true
gradient_checkpointing: true

# ── DeepSpeed ZeRO Stage 2 ──
deepspeed: ds_stage2.json

# ── Logging ──
logging_steps: 10
save_strategy: steps
save_steps: 50

model-out

This model is a fine-tuned version of allenai/Olmo-3-1025-7B on the data/1b/all.jsonl dataset.

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: 5e-05
• train_batch_size: 4
• eval_batch_size: 4
• seed: 42
• distributed_type: multi-GPU
• num_devices: 8
• gradient_accumulation_steps: 4
• total_train_batch_size: 128
• total_eval_batch_size: 32
• optimizer: Use OptimizerNames.ADAMW_TORCH 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: 200
• training_steps: 150

Training results

Framework versions

• Transformers 5.3.0
• Pytorch 2.8.0+cu126
• Datasets 4.5.0
• Tokenizers 0.22.2