Image-Text-to-Text
Transformers
Safetensors
inkling2_loop
weight-edit
browser-automation
conversational
8-bit precision
modelopt
Instructions to use husky8/Inkling-2-Mini-Fast-Browser-Use with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Transformers
How to use husky8/Inkling-2-Mini-Fast-Browser-Use with Transformers:
# Use a pipeline as a high-level helper from transformers import pipeline pipe = pipeline("image-text-to-text", model="husky8/Inkling-2-Mini-Fast-Browser-Use") messages = [ { "role": "user", "content": [ {"type": "image", "url": "https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/p-blog/candy.JPG"}, {"type": "text", "text": "What animal is on the candy?"} ] }, ] pipe(text=messages)# pip install -U transformers accelerate # Load model directly from transformers import AutoProcessor, AutoModelForMultimodalLM processor = AutoProcessor.from_pretrained("husky8/Inkling-2-Mini-Fast-Browser-Use") model = AutoModelForMultimodalLM.from_pretrained("husky8/Inkling-2-Mini-Fast-Browser-Use", device_map="auto") messages = [ { "role": "user", "content": [ {"type": "image", "url": "https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/p-blog/candy.JPG"}, {"type": "text", "text": "What animal is on the candy?"} ] }, ] inputs = processor.apply_chat_template( messages, add_generation_prompt=True, tokenize=True, return_dict=True, return_tensors="pt", ).to(model.device) outputs = model.generate(**inputs, max_new_tokens=256) print(processor.decode(outputs[0][inputs["input_ids"].shape[-1]:])) - Notebooks
- Google Colab
- Kaggle
- Local Apps Settings
- vLLM
How to use husky8/Inkling-2-Mini-Fast-Browser-Use with vLLM:
Install from pip and serve model
# Install vLLM from pip: pip install vllm # Start the vLLM server: vllm serve "husky8/Inkling-2-Mini-Fast-Browser-Use" # Call the server using curl (OpenAI-compatible API): curl -X POST "http://localhost:8000/v1/chat/completions" \ -H "Content-Type: application/json" \ --data '{ "model": "husky8/Inkling-2-Mini-Fast-Browser-Use", "messages": [ { "role": "user", "content": [ { "type": "text", "text": "Describe this image in one sentence." }, { "type": "image_url", "image_url": { "url": "https://cdn.britannica.com/61/93061-050-99147DCE/Statue-of-Liberty-Island-New-York-Bay.jpg" } } ] } ] }'Use Docker
docker model run hf.co/husky8/Inkling-2-Mini-Fast-Browser-Use
- SGLang
How to use husky8/Inkling-2-Mini-Fast-Browser-Use 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 "husky8/Inkling-2-Mini-Fast-Browser-Use" \ --host 0.0.0.0 \ --port 30000 # Call the server using curl (OpenAI-compatible API): curl -X POST "http://localhost:30000/v1/chat/completions" \ -H "Content-Type: application/json" \ --data '{ "model": "husky8/Inkling-2-Mini-Fast-Browser-Use", "messages": [ { "role": "user", "content": [ { "type": "text", "text": "Describe this image in one sentence." }, { "type": "image_url", "image_url": { "url": "https://cdn.britannica.com/61/93061-050-99147DCE/Statue-of-Liberty-Island-New-York-Bay.jpg" } } ] } ] }'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 "husky8/Inkling-2-Mini-Fast-Browser-Use" \ --host 0.0.0.0 \ --port 30000 # Call the server using curl (OpenAI-compatible API): curl -X POST "http://localhost:30000/v1/chat/completions" \ -H "Content-Type: application/json" \ --data '{ "model": "husky8/Inkling-2-Mini-Fast-Browser-Use", "messages": [ { "role": "user", "content": [ { "type": "text", "text": "Describe this image in one sentence." }, { "type": "image_url", "image_url": { "url": "https://cdn.britannica.com/61/93061-050-99147DCE/Statue-of-Liberty-Island-New-York-Bay.jpg" } } ] } ] }' - Docker Model Runner
How to use husky8/Inkling-2-Mini-Fast-Browser-Use with Docker Model Runner:
docker model run hf.co/husky8/Inkling-2-Mini-Fast-Browser-Use
Download benchmark/verify_browser_timings.py from husky8/Inkling-2-Mini-Fast-Browser-Use: direct link, hf CLI and curl.
- Browser
- Download file 3.56 kB
-
https://huggingface.co/husky8/Inkling-2-Mini-Fast-Browser-Use/resolve/main/benchmark/verify_browser_timings.py
- Command line
-
hf download hf://husky8/Inkling-2-Mini-Fast-Browser-Use/benchmark/verify_browser_timings.py
-
curl -L -o verify_browser_timings.py https://huggingface.co/husky8/Inkling-2-Mini-Fast-Browser-Use/resolve/main/benchmark/verify_browser_timings.py
3.56 kB
| """Check every full-task speed ratio and token count against the published episode timings.""" | |
| import argparse | |
| import json | |
| import math | |
| from pathlib import Path | |
| from zipfile import ZipFile | |
| def verify(archive_path, pairs_path): | |
| def check(condition, message): | |
| if not condition: | |
| raise ValueError(message) | |
| pairs=json.loads(pairs_path.read_text())['pairs'] | |
| with ZipFile(archive_path) as archive: | |
| names=[n for n in archive.namelist() if n.startswith('timing/')] | |
| episodes={} | |
| for name in names: | |
| event=json.loads(archive.read(name)) | |
| grade=json.loads(archive.read(name.replace('timing/','receipts/',1))) | |
| check(event['seconds']==grade['elapsed_s'] and event['score']==grade['score'],f'{name}: grade and timing disagree') | |
| times=[s['start_s'] for s in event['steps']]+[event['seconds']] | |
| check(bool(event['steps']) and all(math.isfinite(t) and t>=0 for t in times),f'{name}: invalid timestamps') | |
| for start,end,s in zip(times,times[1:],event['steps']): | |
| check(0<=s['inference_s']<=end-start+1e-6 and 0<=s['thinking']<=s['generated'],f'{name}: invalid step duration or token count') | |
| digest=event['receipt_sha256'] | |
| check(digest not in episodes or episodes[digest]==event,f'{name}: conflicting episode identity') | |
| episodes[digest]=event | |
| used=set() | |
| for p in pairs: | |
| original,edited=(episodes[p[a]['receipt_sha256']] for a in ('original','edited')) | |
| label=f"WebArena {p['task']}, {p['checkpoint']}, repeat {p['repeat']}" | |
| check(original['steps'][0]['view_sha256']==edited['steps'][0]['view_sha256'],f'{label}: different initial observations') | |
| for arm,event in [('original',original),('edited',edited)]: | |
| result=p[arm] | |
| check((p['task'],arm,result['seconds'],result['score'],result['shard_sha256'])==(event['task'],event['arm'],event['seconds'],event['score'],event['shard_sha256']),f'{label}: wrong {arm} episode') | |
| check([s['action'] for s in event['steps']]==result['commands'],f'{label}: different {arm} actions') | |
| check(sum(s['generated'] for s in event['steps'])==result['generated'],f'{label}: different {arm} token count') | |
| check(sum(s['thinking'] for s in event['steps'])==result['thinking'],f'{label}: different {arm} reasoning token count') | |
| check(math.isclose(sum(s['inference_s'] for s in event['steps']),result['inference_seconds'],rel_tol=1e-12),f'{label}: different {arm} inference time') | |
| used.add(event['receipt_sha256']) | |
| if p['comparable'] and original['score']==edited['score']==1: | |
| check(edited['seconds']>0 and isinstance(p['ratio'],(int,float)) and math.isclose(original['seconds']/edited['seconds'],p['ratio'],rel_tol=1e-12),f'{label}: speed ratio disagrees with the clock') | |
| else: | |
| check(p['ratio'] is None,f'{label}: speed ratio on an unsuccessful or incomparable pair') | |
| check(bool(pairs) and used==set(episodes) and len(names)==2*len(pairs),'Incomplete timing archive') | |
| return len(pairs),len(names) | |
| if __name__=='__main__': | |
| parser=argparse.ArgumentParser(description=__doc__) | |
| parser.add_argument('archive',type=Path) | |
| parser.add_argument('pairs',type=Path) | |
| args=parser.parse_args() | |
| count,total=verify(args.archive,args.pairs) | |
| print(f'Verified {count} paired comparisons against {total} recorded episode timings.') | |