Extends Unsloth to add a Ollama API Proxy server to allow use of tools that don't support Unsloth's native OpenAI API
-
Updated
May 19, 2026 - Python
#
unsloth-api
Here are 4 public repositories matching this topic...
Llama-3.2-3B-Instruct distilled on DeepSeek-R1 reasoning traces to think out loud. GSM8K accuracy falls from 75.0% to 65.0% while output length more than doubles — published as a measured negative result, with the per-problem breakdown behind the headline number.
llama lora reasoning fine-tuning distillation peft distillation-model fine-tuning-nlp reasoning-agent fine-tuning-llm llama2 lora-training unsloth reasoning-language-models reasoning-models lora-fine-tuning peft-fine-tuning unsloth-api peft-lora unsloth-studio
-
Updated
Oct 26, 2026 - Jinja
Qwen2.5-3B-Instruct turned into a SQL writer with a 16-bit LoRA. AST-normalised exact match rises from 7.7% to 74.3% on 300 held-out queries, and every output parses as valid SQL. The same cleaning was applied to both models, so the gain is the SQL being right, not the formatting.
sql code-generation sql-generation lora fine-tuning peft text-to-sql text2sql code-generation-techniques fine-tuning-llm lora-training qwen unsloth qwen2-5 text2sql-evaluation text-to-sql-translation text-to-sql-generation code-generation-ai-workflow unsloth-api peft-lora
-
Updated
Oct 26, 2026 - Jinja
Proxy Server that allows Ollama API calls to operate Unsloth
-
Updated
May 17, 2026 - Python
Add this topic to your repo
To associate your repository with the unsloth-api topic, visit your repo's landing page and select "manage topics."