OnAnOrange
/

Dream-7B-Instruct-s1k-sft

@@ -6,201 +6,150 @@ tags:
 - lora
 - transformers
 ---
-# 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. -->
-- **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]
-### Framework versions
-- PEFT 0.17.1

 - lora
 - transformers
 ---
+<center> <div style="text-align: center;"> <img src="https://raw.githubusercontent.com/ZHZisZZ/dllm/main/assets/logo.gif" width="400" />
+ </div> </center>
+# Dream-7B-Instruct-s1k-sft
+Dream-7B-Instruct-s1k-sft is a diffusion-based instruct model post-trained from [Dream-v0-Instruct-7B](https://huggingface.co/Dream-org/Dream-v0-Instruct-7B) on **[simplescaling/s1K](https://huggingface.co/datasets/simplescaling/s1K)**, using [MDLM](https://arxiv.org/abs/2406.07524) (masked diffusion) and trained with the [dLLM](https://github.com/ZHZisZZ/dllm) framework.
+## Model Overview
+Dream-7B-Instruct-s1k-sft has the following features:
+- **Method:** [Masked Diffusion Language Modeling (MDLM)](https://arxiv.org/abs/2406.07524);
+- **Framework:** [dLLM](https://github.com/ZHZisZZ/dllm)
+- **Base model:** [Dream-v0-Instruct-7B](https://huggingface.co/Dream-org/Dream-v0-Instruct-7B)
+- **Dataset (SFT):** [simplescaling/s1K](https://huggingface.co/datasets/simplescaling/s1K)
+For broader training and ablation reporting in the dLLM ecosystem, see the [dLLM paper](https://arxiv.org/pdf/2602.22661).
+Eval notes: Metrics use confidence-threshold decoding (`alg: confidence_threshold`). The primary table is at `confidence_threshold` = 0.9; full grids sweep `confidence_threshold` ∈ {0.6, 0.7, 0.8, 0.9} with `max_new_tokens` ∈ {256, 512}.
+---
+## Primary results
+<table>
+<thead>
+<tr>
+<th align="left"><b>Benchmark</b></th>
+<th align="center"><b>max_new_tokens=256</b><br><sub>(Acc % | TPS)</sub></th>
+<th align="center"><b>max_new_tokens=512</b><br><sub>(Acc % | TPS)</sub></th>
+</tr>
+</thead>
+<tbody>
+<tr>
+<td align="left">GSM8K</td>
+<td align="center">81.80 | 2.30</td>
+<td align="center">84.31 | 2.56</td>
+</tr>
+<tr>
+<td align="left">HumanEval</td>
+<td align="center">54.27 | 2.53</td>
+<td align="center">53.05 | 3.24</td>
+</tr>
+<tr>
+<td align="left">MBPP</td>
+<td align="center">57.80 | 2.21</td>
+<td align="center">57.80 | 2.37</td>
+</tr>
+<tr>
+<td align="left">MATH</td>
+<td align="center">45.16 | 2.32</td>
+<td align="center">49.70 | 2.57</td>
+</tr>
+</tbody>
+</table>
+---
+## Threshold sweep
+<table>
+<thead>
+<tr>
+<th align="left"><b>Benchmark</b></th>
+<th align="center"><b>τ=0.6</b><br><sub>Acc | TPS</sub></th>
+<th align="center"><b>τ=0.7</b><br><sub>Acc | TPS</sub></th>
+<th align="center"><b>τ=0.8</b><br><sub>Acc | TPS</sub></th>
+<th align="center"><b>τ=0.9</b><br><sub>Acc | TPS</sub></th>
+</tr>
+</thead>
+<tbody>
+<tr>
+<td align="left">GSM8K</td>
+<td align="center">65.66 | 4.14</td>
+<td align="center">74.68 | 3.38</td>
+<td align="center">79.83 | 2.66</td>
+<td align="center">81.80 | 2.30</td>
+</tr>
+<tr>
+<td align="left">HumanEval</td>
+<td align="center">34.15 | 4.01</td>
+<td align="center">43.90 | 3.53</td>
+<td align="center">51.83 | 3.00</td>
+<td align="center">54.27 | 2.53</td>
+</tr>
+<tr>
+<td align="left">MBPP</td>
+<td align="center">41.80 | 3.53</td>
+<td align="center">49.40 | 2.71</td>
+<td align="center">55.80 | 2.36</td>
+<td align="center">57.80 | 2.21</td>
+</tr>
+<tr>
+<td align="left">MATH</td>
+<td align="center">37.02 | 3.64</td>
+<td align="center">41.80 | 3.12</td>
+<td align="center">44.48 | 2.67</td>
+<td align="center">45.16 | 2.32</td>
+</tr>
+</tbody>
+</table>
+<code>max_new_tokens=256</code>, columns sweep <code>confidence_threshold</code> ∈ {0.6, 0.7, 0.8, 0.9}
+<table>
+<thead>
+<tr>
+<th align="left"><b>Benchmark</b></th>
+<th align="center"><b>τ=0.6</b><br><sub>Acc | TPS</sub></th>
+<th align="center"><b>τ=0.7</b><br><sub>Acc | TPS</sub></th>
+<th align="center"><b>τ=0.8</b><br><sub>Acc | TPS</sub></th>
+<th align="center"><b>τ=0.9</b><br><sub>Acc | TPS</sub></th>
+</tr>
+</thead>
+<tbody>
+<tr>
+<td align="left">GSM8K</td>
+<td align="center">67.55 | 6.45</td>
+<td align="center">75.44 | 5.34</td>
+<td align="center">81.65 | 3.54</td>
+<td align="center">84.31 | 2.56</td>
+</tr>
+<tr>
+<td align="left">HumanEval</td>
+<td align="center">31.10 | 4.42</td>
+<td align="center">45.73 | 3.98</td>
+<td align="center">50.00 | 3.54</td>
+<td align="center">53.05 | 3.24</td>
+</tr>
+<tr>
+<td align="left">MBPP</td>
+<td align="center">42.20 | 5.60</td>
+<td align="center">50.20 | 3.20</td>
+<td align="center">56.80 | 2.59</td>
+<td align="center">57.80 | 2.37</td>
+</tr>
+<tr>
+<td align="left">MATH</td>
+<td align="center">38.72 | 4.68</td>
+<td align="center">44.10 | 3.90</td>
+<td align="center">48.22 | 3.13</td>
+<td align="center">49.70 | 2.57</td>
+</tr>
+</tbody>
+</table>
+<code>max_new_tokens=512</code>, columns sweep <code>confidence_threshold</code> ∈ {0.6, 0.7, 0.8, 0.9}