Refine Model Card: QServe Information and Links

This PR updates the model card to accurately reflect the QServe model and its capabilities, replacing the outdated VILA information. The `pipeline_tag` is updated to `text-generation`, and key features from the paper and README are summarized.

README.md

@@ -1,37 +1,37 @@
 ---
-license: cc-by-nc-4.0
 library_name: transformers
+license: cc-by-nc-4.0
 pipeline_tag: text-generation
 tags:
-- VILA
-- VLM
+- quantization
+- llm-serving
+- efficient-inference
 ---
 
-# VILA Model Card
+# QServe: W4A8KV4 Quantization and System Co-design for Efficient LLM Serving
 
-## Model details
+This repository contains the QServe inference library and quantized model checkpoints as described in [QServe: W4A8KV4 Quantization and System Co-design for Efficient LLM Serving](https://huggingface.co/papers/2405.04532).  QServe significantly improves the throughput of large language model serving on GPUs through W4A8KV4 quantization and system co-design.
 
-**Model type:**
-VILA is a visual language model (VLM) pretrained with interleaved image-text data at scale, enabling multi-image VLM. VILA is deployable on the edge, including Jetson Orin and laptop by AWQ 4bit quantization through TinyChat framework. We find: (1) image-text pairs are not enough, interleaved image-text is essential; (2) unfreezing LLM during interleaved image-text pre-training enables in-context learning; (3)re-blending text-only instruction data is crucial to boost both VLM and text-only performance. VILA unveils appealing capabilities, including: multi-image reasoning, in-context learning, visual chain-of-thought, and better world knowledge.
+Code: https://github.com/mit-han-lab/qserve
 
-**Model date:**
-VILA1.5-13b was trained in May 2024.
+## Key Features
 
-**Paper or resources for more information:**
-https://github.com/NVLabs/VILA
+QServe offers:
 
-```
-@misc{lin2023vila,
-      title={VILA: On Pre-training for Visual Language Models},
-      author={Ji Lin and Hongxu Yin and Wei Ping and Yao Lu and Pavlo Molchanov and Andrew Tao and Huizi Mao and Jan Kautz and Mohammad Shoeybi and Song Han},
-      year={2023},
-      eprint={2312.07533},
-      archivePrefix={arXiv},
-      primaryClass={cs.CV}
-}
-```
+* **High Throughput:** Achieves significant speedups compared to TensorRT-LLM, especially on cost-effective GPUs like the L40S.  Improves maximum achievable throughput of Llama-3-8B by 1.2x on A100, 1.4x on L40S; and Qwen1.5-72B by 2.4x on A100, 3.5x on L40S.
+* **W4A8KV4 Quantization:** Employs 4-bit weights, 8-bit activations, and 4-bit KV cache for efficient inference.
+* **QoQ Quantization Algorithm:** Implements the quattuor-octo-quattuor (4-8-4) quantization algorithm.
+* **Low Dequantization Overhead:** Minimizes runtime overhead associated with dequantization using progressive quantization.
+* **SmoothAttention:** Mitigates accuracy degradation from 4-bit KV quantization.
+* **Compute-Aware Weight Reordering and Register-Level Parallelism:** Reduces dequantization latency.
+* **Memory-Bound Fused Attention:**  Leverages the performance gains of KV4 quantization.
+* **PyTorch-Based:** Provides PyTorch-level flexibility while maintaining TensorRT-LLM-level efficiency.
+* **Full Support for In-flight Batching and Paged Attention:** Enables efficient handling of large requests and long sequences.
+* **Large-Scale Generation Support:** Enables efficient large-scale synthetic data generation.
+
+
+## Citation
 
-https://github.com/mit-han-lab/qserve
 ```
 @article{lin2024qserve,
   title={QServe: W4A8KV4 Quantization and System Co-design for Efficient LLM Serving},
@@ -39,86 +39,4 @@ https://github.com/mit-han-lab/qserve
   journal={arXiv preprint arXiv:2405.04532},
   year={2024}
 }
-```
-
-## License
-- The code is released under the Apache 2.0 license as found in the [LICENSE](./LICENSE) file.
-- The pretrained weights are released under the [CC-BY-NC-SA-4.0 license](https://creativecommons.org/licenses/by-nc-sa/4.0/deed.en).
-- The service is a research preview intended for non-commercial use only, and is subject to the following licenses and terms:
-    - [Model License](https://github.com/facebookresearch/llama/blob/main/MODEL_CARD.md) of LLaMA
-    - [Terms of Use](https://openai.com/policies/terms-of-use) of the data generated by OpenAI
-    - [Dataset Licenses](https://github.com/Efficient-Large-Model/VILA/blob/main/data_prepare/LICENSE) for each one used during training.
-
-**Where to send questions or comments about the model:**
-https://github.com/NVLabs/VILA/issues
-
-## Intended use
-**Primary intended uses:**
-The primary use of VILA is research on large multimodal models and chatbots.
-
-**Primary intended users:**
-The primary intended users of the model are researchers and hobbyists in computer vision, natural language processing, machine learning, and artificial intelligence.
-
-## Model Architecture: 
-**Architecture Type:** Transformer
-**Network Architecture:** siglip, vicuna1.5
-
-## Input:
-**Input Type:** Image, Video, Text
-**Input Format:** Red, Green, Blue; MP4 ;String
-**Input Parameters:** 2D, 3D
-
-## Output: 
-**Output Type:** Text
-**Output Format:** String
-
-**Supported Hardware Microarchitecture Compatibility:**
-* Ampere
-* Jetson
-* Hopper
-* Lovelace
-
-**[Preferred/Supported] Operating System(s):** <br>
-Linux
-
-## Model Version(s):
-* VILA1.5-3B
-* VILA1.5-3B-s2
-* Llama-3-VILA1.5-8B
-* VILA1.5-13B
-* VILA1.5-40B
-* VILA1.5-3B-AWQ
-* VILA1.5-3B-s2-AWQ
-* Llama-3-VILA1.5-8B-AWQ
-* VILA1.5-13B-AWQ
-* VILA1.5-40B-AWQ
-
-## Training dataset
-See [Dataset Preparation](https://github.com/NVLabs/VILA/blob/main/data_prepare/README.md) for more details.
-
-** Data Collection Method by dataset
-* [Hybrid: Automated, Human]
-
-** Labeling Method by dataset
-* [Hybrid: Automated, Human]
-
-**Properties (Quantity, Dataset Descriptions, Sensor(s)):**
-53 million image-text pairs or interleaved image text content.
-
-
-## Evaluation dataset
-A collection of 12 benchmarks, including 5 academic VQA benchmarks and 7 recent benchmarks specifically proposed for instruction-following LMMs.
-
-## Inference:
-**Engine:** [Tensor(RT), Triton, Or List Other Here]
-* PyTorch
-* TensorRT-LLM
-* TinyChat
-
-**Test Hardware:**
-* A100
-* Jetson Orin
-* RTX 4090
-
-## Ethical Considerations
-NVIDIA believes Trustworthy AI is a shared responsibility and we have established policies and practices to enable development for a wide array of AI applications.  When downloaded or used in accordance with our terms of service, developers should work with their internal model team to ensure this model meets requirements for the relevant industry and use case and addresses unforeseen product misuse.  
+```