Update README.md
Browse files
README.md
CHANGED
|
@@ -1,3 +1,76 @@
|
|
| 1 |
-
---
|
| 2 |
-
license: apache-2.0
|
| 3 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
license: apache-2.0
|
| 3 |
+
language:
|
| 4 |
+
- en
|
| 5 |
+
tags:
|
| 6 |
+
- code
|
| 7 |
+
- sarcasm
|
| 8 |
+
- chandler-bing
|
| 9 |
+
- lora
|
| 10 |
+
- transformers
|
| 11 |
+
metrics:
|
| 12 |
+
- code_eval
|
| 13 |
+
pipeline_tag: text-generation
|
| 14 |
+
base_model: dgtalbug/stable-code-instruct-3b-base
|
| 15 |
+
---
|
| 16 |
+
|
| 17 |
+
# Stephen — Sarcastic Coding Assistant
|
| 18 |
+
|
| 19 |
+

|
| 20 |
+
|
| 21 |
+
## Model Description
|
| 22 |
+
|
| 23 |
+
**Stephen** is a fine-tuned variant of `stable-code-instruct-3b` with a personality inspired by:
|
| 24 |
+
- **Chandler Bing** (*Friends*) — sarcastic wit
|
| 25 |
+
- **Deadpool** — meta humor & breaking the fourth wall
|
| 26 |
+
- **Senior Dev energy** — opinionated code roasting
|
| 27 |
+
|
| 28 |
+
Stephen is trained on:
|
| 29 |
+
- *Friends* transcripts (dialogue style)
|
| 30 |
+
- Reddit jokes datasets
|
| 31 |
+
- Sarcasm headlines
|
| 32 |
+
- Coding & programming humor datasets
|
| 33 |
+
|
| 34 |
+
### Intended Use
|
| 35 |
+
- Writing sarcastic code comments
|
| 36 |
+
- Generating humorous coding explanations
|
| 37 |
+
- Adding playful banter to code reviews
|
| 38 |
+
- Conversational AI with a strong personality
|
| 39 |
+
|
| 40 |
+
⚠ **Not for serious enterprise documentation unless you enjoy snarky footnotes.**
|
| 41 |
+
|
| 42 |
+
## Training Details
|
| 43 |
+
- **Base Model**: `dgtalbug/stable-code-instruct-3b-base`
|
| 44 |
+
- **Fine-tuning Method**: LoRA + PEFT
|
| 45 |
+
- **Framework**: Transformers, BitsAndBytes
|
| 46 |
+
- **Datasets**: Friends transcripts, Reddit jokes, Sarcasm headlines, Programming humor
|
| 47 |
+
|
| 48 |
+
## Example Usage
|
| 49 |
+
|
| 50 |
+
```python
|
| 51 |
+
from transformers import AutoTokenizer, AutoModelForCausalLM
|
| 52 |
+
import torch
|
| 53 |
+
|
| 54 |
+
model_id = "dgtalbug/stephen"
|
| 55 |
+
|
| 56 |
+
tokenizer = AutoTokenizer.from_pretrained(model_id)
|
| 57 |
+
model = AutoModelForCausalLM.from_pretrained(model_id, torch_dtype=torch.bfloat16).eval()
|
| 58 |
+
|
| 59 |
+
prompt = "Explain bubble sort as if I am a junior dev who just broke production."
|
| 60 |
+
inputs = tokenizer(prompt, return_tensors="pt").to(model.device)
|
| 61 |
+
outputs = model.generate(**inputs, max_new_tokens=150)
|
| 62 |
+
print(tokenizer.decode(outputs[0], skip_special_tokens=True))
|
| 63 |
+
```
|
| 64 |
+
|
| 65 |
+
---
|
| 66 |
+
|
| 67 |
+
## Citation
|
| 68 |
+
|
| 69 |
+
```bibtex
|
| 70 |
+
@misc{stephen-sarcastic-coder,
|
| 71 |
+
title = {Stephen: Sarcastic Coding Assistant},
|
| 72 |
+
author = {dgtalbug},
|
| 73 |
+
year = {2025},
|
| 74 |
+
howpublished = {\url{https://huggingface.co/dgtalbug/stephen}}
|
| 75 |
+
}
|
| 76 |
+
```
|