Introduction of doctomodernbert-fr-large-diffusion-v0.1
Model Details of doctomodernbert-fr-large-diffusion-v0.1
DoctoModernBERT-fr-large-diffusion-v0.1
A text diffusion model based on
DoctoModernBERT-fr-large
, finetuned on a medical instruction-following dataset.
Unlike an autoregressive LLM, which decodes strictly left to right one token at a time, this model works by iterative denoising. It starts from a fully masked canvas and, at each step, predicts all masked positions at once and keeps only the most confident ones. It still moves through the sequence one block at a time, so generation is parallel inside each block but left to right across blocks.
This is an exploratory checkpoint. It's small and lightly trained, so expect mistakes and weaker answers.
🚀 How to Use
Training and generation are done through the
dLLM
library.
Setup
git clone https://github.com/ZHZisZZ/dllm
cd dllm
pip install -e .
Sampling
import dllm
model_id = "bofenghuang/doctomodernbert-fr-large-diffusion-v0.1"# or a local path
model = dllm.utils.get_model(model_name_or_path=model_id).eval()
tokenizer = dllm.utils.get_tokenizer(model_name_or_path=model_id)
sampler = dllm.core.samplers.MDLMSampler(model=model, tokenizer=tokenizer)
config = dllm.core.samplers.MDLMSamplerConfig(
steps=128, # total diffusion steps
max_new_tokens=128, # length of the generated answer span
block_size=32, # semi-autoregressive block size
temperature=0.0, # 0.0 = greedy / low-confidence remasking
remasking="low_confidence",
)
messages = [
[{"role": "user", "content": "Quels sont les symptômes typiques d'une infection urinaire ?"}],
]
inputs = tokenizer.apply_chat_template(messages, add_generation_prompt=True, tokenize=True)
outputs = sampler.sample(inputs, config, return_dict=True)
print(dllm.utils.sample_trim(tokenizer, outputs.sequences.tolist(), inputs)[0])
doctomodernbert-fr-large-diffusion-v0.1 huggingface.co is an AI model on huggingface.co that provides doctomodernbert-fr-large-diffusion-v0.1's model effect (), which can be used instantly with this bofenghuang doctomodernbert-fr-large-diffusion-v0.1 model. huggingface.co supports a free trial of the doctomodernbert-fr-large-diffusion-v0.1 model, and also provides paid use of the doctomodernbert-fr-large-diffusion-v0.1. Support call doctomodernbert-fr-large-diffusion-v0.1 model through api, including Node.js, Python, http.
doctomodernbert-fr-large-diffusion-v0.1 huggingface.co is an online trial and call api platform, which integrates doctomodernbert-fr-large-diffusion-v0.1's modeling effects, including api services, and provides a free online trial of doctomodernbert-fr-large-diffusion-v0.1, you can try doctomodernbert-fr-large-diffusion-v0.1 online for free by clicking the link below.
bofenghuang doctomodernbert-fr-large-diffusion-v0.1 online free url in huggingface.co:
doctomodernbert-fr-large-diffusion-v0.1 is an open source model from GitHub that offers a free installation service, and any user can find doctomodernbert-fr-large-diffusion-v0.1 on GitHub to install. At the same time, huggingface.co provides the effect of doctomodernbert-fr-large-diffusion-v0.1 install, users can directly use doctomodernbert-fr-large-diffusion-v0.1 installed effect in huggingface.co for debugging and trial. It also supports api for free installation.
doctomodernbert-fr-large-diffusion-v0.1 install url in huggingface.co: