micic-mihajlo/gemma-4-12b-social-post-lora

micic-mihajlo/gemma-4-12b-social-post-lora

LoRA adapter for general-purpose social media post drafting.

Base

• —Base model: google/gemma-4-12B
• —Adapter: micic-mihajlo/gemma-4-12b-social-post-lora
• —Training method: QLoRA + FSDP + MMD/JMQ

Data

• —Train examples: 3
• —Validation examples: 1
• —Generated eval posts: 10

Objective

The adapter is optimized for brief-to-post generation with supervised cross-entropy plus distribution-alignment regularizers over token-logit shape.

Evaluation

Base Comparison

• —Style distance improvement: 661.4012
• —Adapter-target distance: 282.1723
• —Base-target distance: 943.5735

Limitations

• —This adapter should not be used to imitate a private person, brand, or protected identity.
• —Quality depends on the human-reference corpus and held-out brief coverage.
• —Distribution metrics are diagnostic signals, not a substitute for human review.