Instructions to use caikybaldo999/webcoder-100m-html-css-js with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use caikybaldo999/webcoder-100m-html-css-js with Transformers:

# Use a pipeline as a high-level helper
from transformers import pipeline

pipe = pipeline("text-generation", model="caikybaldo999/webcoder-100m-html-css-js")

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("caikybaldo999/webcoder-100m-html-css-js")
model = AutoModelForCausalLM.from_pretrained("caikybaldo999/webcoder-100m-html-css-js", device_map="auto")

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use caikybaldo999/webcoder-100m-html-css-js with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "caikybaldo999/webcoder-100m-html-css-js"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "caikybaldo999/webcoder-100m-html-css-js",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Use Docker

docker model run hf.co/caikybaldo999/webcoder-100m-html-css-js

SGLang

How to use caikybaldo999/webcoder-100m-html-css-js with SGLang:

Install from pip and serve model

# Install SGLang from pip:
pip install sglang
# Start the SGLang server:
python3 -m sglang.launch_server \
    --model-path "caikybaldo999/webcoder-100m-html-css-js" \
    --host 0.0.0.0 \
    --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "caikybaldo999/webcoder-100m-html-css-js",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Use Docker images

docker run --gpus all \
    --shm-size 32g \
    -p 30000:30000 \
    -v ~/.cache/huggingface:/root/.cache/huggingface \
    --env "HF_TOKEN=<secret>" \
    --ipc=host \
    lmsysorg/sglang:latest \
    python3 -m sglang.launch_server \
        --model-path "caikybaldo999/webcoder-100m-html-css-js" \
        --host 0.0.0.0 \
        --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "caikybaldo999/webcoder-100m-html-css-js",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Docker Model Runner
How to use caikybaldo999/webcoder-100m-html-css-js with Docker Model Runner:
```
docker model run hf.co/caikybaldo999/webcoder-100m-html-css-js
```

webcoder-100m-html-css-js / README.md

caikybaldo999

Upload folder using huggingface_hub

ee9c8c2 verified 1 day ago

preview code

Raw

History Blame Contribute Delete

1.4 kB

	---
	language:
	- code
	library_name: transformers
	pipeline_tag: text-generation
	tags:
	- html
	- css
	- javascript
	- code
	- llama
	- from-scratch
	---

	# WebCoder-100M

	A small decoder-only model specialized in HTML, CSS and JavaScript.

	## Architecture
	- Parameters: 99,894,528
	- Layers: 11
	- Hidden size: 768
	- Attention heads: 12
	- Vocabulary: 28,672
	- Max context: 2,048
	- Training sequence length: 1,024

	## Training
	The model was initialized from scratch.

	1. Causal pre-training on HTML/CSS/JavaScript.
	2. Instruction fine-tuning on web-development instruction/code pairs.

	## Token accounting
	- Total processed: 582,209,140
	- Pre-training: 568,246,272
	- SFT processed: 13,962,868
	- SFT supervised response tokens: 9,124,446
	- Global cap: 2,000,000,000

	## Data
	Pre-training: `bigcode/the-stack-smol-xl`, HTML/JavaScript/CSS subsets.

	Instruction tuning: `iamtarun/code_instructions_120k_alpaca`, filtered for web-development examples.

	Review upstream dataset cards and source licenses before commercial use.

	## Prompt format
	```text
	<\|system\|>
	You are WebCoder...<\|end\|>
	<\|user\|>
	Create a responsive landing page...<\|end\|>
	<\|assistant\|>
	...
	```

	## Limitations
	This is a roughly 100M-parameter model trained from scratch. Its quality depends
	strongly on how many tokens were actually processed. Generated code can contain
	bugs or security issues and should be reviewed.