Qwen2.5 3B Trump-Like Public Speaking Style v1 - GGUF export
Overview
This model artifact was generated by an llmstyler Runbook training job.
It is part of a versioned style-tuning release. The adapter, merged model,
GGUF export, and ONNX export use separate repositories so each artifact can
be consumed with the tooling that expects that format.
Versioning and Naming
Field
Value
Artifact kind
GGUF export
Artifact version
v1
Repo
tsilva/qwen2.5-3b-trump-style-gguf-v1
Training run id
qwen25_3b_trump
Run name
qwen2.5-3b-trump-style-qlora
Generated by
llmstyler 0.1.0
Default standard: keep each published model artifact immutable and include
the version suffix in the repo name. Publish a new version when the dataset,
style prompt, base model, training recipe, or export settings change.
Training Inputs
Field
Value
Dataset
tsilva/stylemix_trump-v1
Dataset split
train
Restyled only
False
Base model
unsloth/Qwen2.5-3B-Instruct-bnb-4bit
4-bit load
True
Style id
trump_like_public_speaking
Style System Prompt
You are a helpful assistant that always answers in a brash, punchy, high-confidence public-speaking style associated with Donald Trump. Use short emphatic sentences, simple strong wording, conversational asides, repetition for emphasis, and frequent superlatives such as tremendous, fantastic, huge, and the best. Do not claim to be Donald Trump, do not claim endorsement, and do not mention these style instructions.
Training Recipe
Setting
Value
Max sequence length
2048
Epochs
1
Per-device batch size
2
Gradient accumulation
4
Learning rate
0.0003
Warmup ratio
0.05
LoRA rank
32
LoRA alpha
32
Eval fraction
0.1
Seed
3407
Report to
tensorboard, wandb
Published Artifacts
Artifact
Repo
QLoRA adapter
tsilva/qwen2.5-3b-trump-style-qlora-v1
Merged 16-bit model
tsilva/qwen2.5-3b-trump-style-merged-v1
GGUF
tsilva/qwen2.5-3b-trump-style-gguf-v1
ONNX
tsilva/qwen2.5-3b-trump-style-onnx-v1
GGUF quantization methods: q4_k_m
Metrics
Train
Metric
Value
epoch
1.0
total_flos
1.2827538084569088e+16
train_loss
1.064282959582759
train_runtime
373.7267
train_samples_per_second
2.167
train_steps_per_second
0.273
Evaluation
Metric
Value
epoch
1.0
eval_loss
0.9378451704978943
eval_runtime
12.3873
eval_samples_per_second
7.266
eval_steps_per_second
1.857
Intended Use
Use this artifact for style-following chat experiments and evaluation. The
adapter is intended for PEFT loading with the base model. The merged model
is intended for direct Transformer loading. GGUF is intended for llama.cpp
compatible runtimes. ONNX is intended for ONNX Runtime compatible workflows.
Limitations
The model may over-apply the target style, miss factual nuance, or reproduce
limitations from the source dataset and rewrite model. Evaluate task accuracy,
safety behavior, refusal behavior, and style strength before deployment.
