---

license: apache-2.0

base_model:

- unsloth/gemma-4-E4B-it

library_name: gguf

pipeline_tag: text-generation

tags:

- gguf

- llama.cpp

- quantized

- unsloth

- lora

- trl

- sft

---

Gemma4-E4B-SFT-Fable5-GGUF

GGUF quantizations of a LoRA fine-tune of unsloth/gemma-4-E4B-it, supervised fine-tuned on ermiaazarkhalili/Fable-5-Complete-2M-Clean (private).

Quantized from ermiaazarkhalili/Gemma4-E4B-SFT-Fable5. See that repository for the full-precision weights.

| | |

| --- | --- |

| Base model | unsloth/gemma-4-E4B-it |

| Training data | ermiaazarkhalili/Fable-5-Complete-2M-Clean (private) |

| Method | LoRA supervised fine-tuning via Unsloth + TRL |

| License | apache-2.0 (inherited from the base model) |

Available quantizations

| File | Size |

| --- | --- |

| gemma4-e4b-sft-fable5.q4_k_m.gguf | 5.34 GB |

| gemma4-e4b-sft-fable5.q5_k_m.gguf | 5.76 GB |

| gemma4-e4b-sft-fable5.q8_0.gguf | 8.03 GB |

Usage

llama.cpp

huggingface-cli download ermiaazarkhalili/Gemma4-E4B-SFT-Fable5-GGUF gemma4-e4b-sft-fable5.q4_k_m.gguf --local-dir .
llama-cli -m gemma4-e4b-sft-fable5.q4_k_m.gguf -p "Explain gradient checkpointing in two sentences." -n 256

Ollama

echo 'FROM ./gemma4-e4b-sft-fable5.q4_k_m.gguf' > Modelfile
ollama create gemma4-e4b-sft-fable5-gguf -f Modelfile
ollama run gemma4-e4b-sft-fable5-gguf

Training configuration

| Setting | Value |

| --- | --- |

| LoRA rank (r) | 16 |

| LoRA alpha | 16 |

| Learning rate | 0.0002 |

| Epochs | 2 |

| Effective batch size | 8 (2 x 4 grad accum) |

| Max sequence length | 4096 |

| Base precision | 4-bit (QLoRA) |

Observed training loss

Measured from our SLURM logs for this configuration. These are training-loss

observations only — no downstream benchmark evaluation has been run on this

model, so they should not be read as a quality claim.

| SLURM job | Steps | First loss | Final loss |

| --- | --- | --- | --- |

| 53221041 | 94,252 | 2.7514 | 0.9215 |

Limitations

• No benchmark evaluation has been run on this checkpoint. The only reported

numbers are training-loss observations.

• Inherits the biases, knowledge cutoff and failure modes of the base model.
• Fine-tuned on a single instruction-following dataset; behaviour outside that

distribution is untested.

• LoRA adapters were merged into the base weights, so the merged model cannot

be detached from this fine-tune.

Reproducing

Trained by notebooks/fable_distillation_gemma4-e4b_fable_unsloth.ipynb, executed non-interactively with

papermill on a SLURM H100 partition (Unsloth + TRL, LoRA).

---

Card generated from the training run's own configuration and logs by

scripts/generate_hub_model_card.py.