llm-model-lab/gemma3-27b-orpo-thinking-v2-checkpoint-1712

Gemma3-27B ORPO Thinking v2 - Checkpoint (Epoch 3)

Epoch 3 checkpoint of ORPO training (final)

Parent Model: llm-model-lab/gemma3-27b-orpo-thinking-v2

Training Metrics (Epoch 3)

Usage with vLLM

vllm serve llm-model-lab/thinking-v2-3epoch \
  --enable-lora \
  --lora-modules epoch3=llm-model-lab/gemma3-27b-orpo-thinking-v2-checkpoint-1712 \
  --max-lora-rank 128 \
  --dtype bfloat16

Files

This checkpoint contains:

• —adapter_model.safetensors: LoRA weights
• —adapter_config.json: PEFT configuration
• —tokenizer.json, tokenizer_config.json: Tokenizer
• —chat_template.jinja: Gemma3 chat template
• —DeepSpeed training state files (for training resume only)