---

language:

• en

library_name: gguf

pipeline_tag: text-generation

license: apache-2.0

base_model: Qwen/Qwen3.6-35B-A3B

base_model_relation: finetune

datasets:

• aisamdasu/algocean-fable5-traces

tags:

• gguf
• llama.cpp
• qwen3.6
• qwen3.5-moe
• lora
• sft
• tool-use
• agentic
• langgraph
• fable5

---

Algocean-Qwen3.6-35B-A3B.gguf

Merged GGUF F16 release of Qwen/Qwen3.6-35B-A3B fine-tuned with a LoRA SFT run on the aisamdasu/algocean-fable5-traces trace mix.

The published GGUF uses checkpoint 3600, which had the lowest validation loss in the training run.

Use

llama-cli -m Algocean-Qwen3.6-35B-A3B.gguf -p "Create a concise LangGraph plan for a repo refactor." -n 512

Artifact

| item | value |

|---|---:|

| Format | GGUF F16 |

| File | Algocean-Qwen3.6-35B-A3B.gguf |

| File size | 71,066,993,920 bytes |

| GGUF version | 3 |

| GGUF metadata keys | 43 |

| GGUF tensors | 753 |

| LoRA tensors merged | 20 |

| SHA-256 | 564448e3c3861fd9f143bba4311c4a3cdc92da6709099fa25d5ed1cc0a4c3e89 |

Training

| key | value |

|---|---|

| Base model | Qwen/Qwen3.6-35B-A3B |

| Dataset | aisamdasu/algocean-fable5-traces |

| Train / eval rows | 30,265 / 512 |

| Selected source rows | Crownelius 26,938 + kelexine 3,839 |

| Dedup skipped | 1,721 |

| Fine-tune method | LoRA SFT |

| LoRA target modules | q_proj, v_proj |

| LoRA rank / alpha / dropout | 16 / 32 / 0.0 |

| Max sequence length | 16,384 |

| Optimizer / schedule | AdamW fused / cosine |

| Learning rate / warmup | 1e-5 / 0.03 |

| Batch / grad accumulation | 1 / 1 |

| Max steps | 4,000 |

| Eval / save interval | 200 / 200 |

| GPU | Modal B200 x1 |

| Runtime | 7,934.71 sec |

Checkpoint Selection

| step | eval loss | eval PPL | note |

|---:|---:|---:|---|

| 200 | 1.300712 | 3.671911 | first eval |

| 3200 | 1.067631 | 2.908480 | near plateau |

| 3400 | 1.067996 | 2.909543 | regression |

| 3600 | 1.066911 | 2.906388 | selected best |

| 3800 | 1.067495 | 2.908084 | worse than best |

| 4000 | 1.067005 | 2.906661 | final, slightly worse |

Best checkpoint rationale:

• 3600 has the lowest validation loss: 1.066911.
• Final step 4000 is very close but still higher: +0.000094 eval loss.
• The curve flattened after 3200, so selecting 3600 keeps the best validation point without chasing extra train loss.

!Eval loss vs step

!Train vs eval loss

!Learning rate vs step

Metrics

| metric | value |

|---|---:|

| Best eval loss | 1.066911 |

| Best eval perplexity | 2.906388 |

| Final eval loss | 1.067005 |

| Final eval perplexity | 2.906661 |

| Trainer average train loss | 0.494604 |

| Last logged train loss | 0.996137 |

| Train samples/sec | 0.504 |

| Train steps/sec | 0.504 |

| Eval samples/sec at final eval | 2.088 |

Comparison Notes

This release is the best checkpoint from the completed Algocean LoRA run. It is not claiming a public benchmark win over the base model or the upstream OptiQ model; those require the same held-out evaluation harness on all models. The selection here is based on the run's validation loss, where checkpoint 3600 is the lowest-loss point among all saved checkpoints.

Intended Use

This model is intended for local llama.cpp-compatible inference experiments around agentic coding, tool-use traces, and LangGraph-style planning. It is not a general safety-tuned assistant release.