Ayushnangia
/

mb-diff-1000step

@@ -1,199 +1,121 @@
 ---
 library_name: transformers
-tags: []
 ---
-# Model Card for Model ID
-<!-- Provide a quick summary of what the model is/does. -->
-## Model Details
-### Model Description
-<!-- Provide a longer summary of what this model is. -->
-This is the model card of a 🤗 transformers model that has been pushed on the Hub. This model card has been automatically generated.
-- **Developed by:** [More Information Needed]
-- **Funded by [optional]:** [More Information Needed]
-- **Shared by [optional]:** [More Information Needed]
-- **Model type:** [More Information Needed]
-- **Language(s) (NLP):** [More Information Needed]
-- **License:** [More Information Needed]
-- **Finetuned from model [optional]:** [More Information Needed]
-### Model Sources [optional]
-<!-- Provide the basic links for the model. -->
-- **Repository:** [More Information Needed]
-- **Paper [optional]:** [More Information Needed]
-- **Demo [optional]:** [More Information Needed]
-## Uses
-<!-- Address questions around how the model is intended to be used, including the foreseeable users of the model and those affected by the model. -->
-### Direct Use
-<!-- This section is for the model use without fine-tuning or plugging into a larger ecosystem/app. -->
-[More Information Needed]
-### Downstream Use [optional]
-<!-- This section is for the model use when fine-tuned for a task, or when plugged into a larger ecosystem/app -->
-[More Information Needed]
-### Out-of-Scope Use
-<!-- This section addresses misuse, malicious use, and uses that the model will not work well for. -->
-[More Information Needed]
-## Bias, Risks, and Limitations
-<!-- This section is meant to convey both technical and sociotechnical limitations. -->
-[More Information Needed]
-### Recommendations
-<!-- This section is meant to convey recommendations with respect to the bias, risk, and technical limitations. -->
-Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
-## How to Get Started with the Model
-Use the code below to get started with the model.
-[More Information Needed]
 ## Training Details
-### Training Data
-<!-- This should link to a Dataset Card, perhaps with a short stub of information on what the training data is all about as well as documentation related to data pre-processing or additional filtering. -->
-[More Information Needed]
-### Training Procedure
-<!-- This relates heavily to the Technical Specifications. Content here should link to that section when it is relevant to the training procedure. -->
-#### Preprocessing [optional]
-[More Information Needed]
-#### Training Hyperparameters
-- **Training regime:** [More Information Needed] <!--fp32, fp16 mixed precision, bf16 mixed precision, bf16 non-mixed precision, fp16 non-mixed precision, fp8 mixed precision -->
-#### Speeds, Sizes, Times [optional]
-<!-- This section provides information about throughput, start/end time, checkpoint size if relevant, etc. -->
-[More Information Needed]
-## Evaluation
-<!-- This section describes the evaluation protocols and provides the results. -->
-### Testing Data, Factors & Metrics
-#### Testing Data
-<!-- This should link to a Dataset Card if possible. -->
-[More Information Needed]
-#### Factors
-<!-- These are the things the evaluation is disaggregating by, e.g., subpopulations or domains. -->
-[More Information Needed]
-#### Metrics
-<!-- These are the evaluation metrics being used, ideally with a description of why. -->
-[More Information Needed]
-### Results
-[More Information Needed]
-#### Summary
-## Model Examination [optional]
-<!-- Relevant interpretability work for the model goes here -->
-[More Information Needed]
-## Environmental Impact
-<!-- Total emissions (in grams of CO2eq) and additional considerations, such as electricity usage, go here. Edit the suggested text below accordingly -->
-Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
-- **Hardware Type:** [More Information Needed]
-- **Hours used:** [More Information Needed]
-- **Cloud Provider:** [More Information Needed]
-- **Compute Region:** [More Information Needed]
-- **Carbon Emitted:** [More Information Needed]
-## Technical Specifications [optional]
-### Model Architecture and Objective
-[More Information Needed]
-### Compute Infrastructure
-[More Information Needed]
-#### Hardware
-[More Information Needed]
-#### Software
-[More Information Needed]
-## Citation [optional]
-<!-- If there is a paper or blog post introducing the model, the APA and Bibtex information for that should go in this section. -->
-**BibTeX:**
-[More Information Needed]
-**APA:**
-[More Information Needed]
-## Glossary [optional]
-<!-- If relevant, include terms and calculations in this section that can help readers understand the model or model card. -->
-[More Information Needed]
-## More Information [optional]
-[More Information Needed]
-## Model Card Authors [optional]
-[More Information Needed]
-## Model Card Contact
-[More Information Needed]

 ---
+license: apache-2.0
+language:
+- en
 library_name: transformers
+tags:
+- modernbert
+- diffusion
+- masked-language-model
+- text-generation
+base_model: answerdotai/ModernBERT-large
+datasets:
+- Ayushnangia/dolci-diffusion-sft-0.9-passrate
+pipeline_tag: fill-mask
 ---
+# mb-diff-1000step
+A ModernBERT-large model fine-tuned as a **diffusion language model** (LLADA-style) for instruction-following.
+## Model Description
+This model uses iterative unmasking for text generation:
+1. Start with a user prompt + fully masked response slots
+2. Model predicts all masked tokens simultaneously
+3. Keep the most confident prediction, repeat until done
+Unlike autoregressive models, this allows parallel token prediction and flexible generation order.
 ## Training Details
+- **Base model**: [answerdotai/ModernBERT-large](https://huggingface.co/answerdotai/ModernBERT-large)
+- **Training data**: [Ayushnangia/dolci-diffusion-sft-0.9-passrate](https://huggingface.co/datasets/Ayushnangia/dolci-diffusion-sft-0.9-passrate) (117k high-quality examples)
+- **Training steps**: 1000
+- **Hardware**: H100 80GB
+- **Variable masking**: 15-99% of assistant tokens masked per sample
+## Usage
+```python
+import torch
+from transformers import AutoTokenizer, AutoModelForMaskedLM
+model_id = "Ayushnangia/mb-diff-1000step"
+device = "cuda" if torch.cuda.is_available() else "cpu"
+tokenizer = AutoTokenizer.from_pretrained(model_id)
+model = AutoModelForMaskedLM.from_pretrained(model_id, torch_dtype=torch.bfloat16).to(device)
+# Build prompt with masked response
+query = "What is 2+2?"
+messages = [{"role": "user", "content": query}]
+prompt = tokenizer.apply_chat_template(messages, tokenize=False, add_generation_prompt=True)
+# Add masked tokens for response
+num_tokens = 64
+prompt_ids = tokenizer.encode(prompt, add_special_tokens=False)
+mask_id = tokenizer.mask_token_id
+im_end_id = tokenizer.convert_tokens_to_ids("<|im_end|>")
+ids = [tokenizer.cls_token_id] + prompt_ids + [mask_id] * num_tokens + [im_end_id]
+# Iterative unmasking
+for step in range(num_tokens):
+    with torch.no_grad():
+        input_tensor = torch.tensor([ids], device=device)
+        logits = model(input_ids=input_tensor).logits[0]
+    probs = torch.softmax(logits, dim=-1)
+    # Find mask positions
+    mask_positions = [i for i, tok in enumerate(ids) if tok == mask_id]
+    if not mask_positions:
+        break
+    # Get confidence for each mask position
+    mask_probs = torch.zeros_like(probs)
+    for pos in mask_positions:
+        mask_probs[pos] = probs[pos]
+    # Fill most confident prediction
+    max_probs, max_tokens = mask_probs.max(dim=-1)
+    best_pos = max_probs.argmax().item()
+    ids[best_pos] = max_tokens[best_pos].item()
+# Decode response
+response_start = len(prompt_ids) + 1
+response_ids = [t for t in ids[response_start:] if t not in (mask_id, im_end_id)]
+response = tokenizer.decode(response_ids, skip_special_tokens=True)
+print(response)
+```
+## Inference Script
+For easier inference, use the sampling script from the training repo:
+```bash
+git clone https://github.com/agokrani/diffu-convert
+cd diffu-convert
+pip install -e .
+python scripts/mb_dllm_sample.py --model Ayushnangia/mb-diff-1000step --query "What is 2+2?"
+```
+## Limitations
+- Early checkpoint (1000 steps) - not fully converged
+- Best for short responses (64-256 tokens)
+- Math/reasoning tasks may have lower accuracy than autoregressive models
+## Citation
+```bibtex
+@misc{mb-diffusion,
+  author = {Ayush Nangia},
+  title = {ModernBERT Diffusion Language Model},
+  year = {2025},
+  publisher = {HuggingFace},
+  url = {https://huggingface.co/Ayushnangia/mb-diff-1000step}
+}
+```