Add pipeline tag and paper link to model card

README.md

@@ -1,9 +1,10 @@
 ---
-license: apache-2.0
-datasets:
-- open-r1/OpenR1-Math-220k
 base_model:
 - Qwen/Qwen3-14B
+datasets:
+- open-r1/OpenR1-Math-220k
+license: apache-2.0
+pipeline_tag: text-generation
 tags:
 - math
 - trimkv
@@ -12,15 +13,18 @@ tags:
 - Compression
 ---
 
-> TRIM-KV is an efficient and learnable key–value eviction strategy designed to improve the efficiency of large language models (LLMs) in long-horizon inference.
+# TrimKV: Token Retention for Memory-Bounded Key-Value Eviction
+
+TRIM-KV is an efficient and learnable key–value eviction strategy designed to improve the efficiency of large language models (LLMs) in long-horizon inference. It was introduced in the paper [Make Each Token Count: Towards Improving Long-Context Performance with KV Cache Eviction](https://huggingface.co/papers/2605.09649) by Ngoc Bui, Hieu Trung Nguyen, Arman Cohan, and Rex Ying.
 
 The core idea behind TRIM-KV is to learn the intrinsic importance of each key–value pair at creation time, which we call *token retention*, and then decay this importance exponentially over time to mimic the standard inference running with eviction. 
 
 The retention score is query-agnostic and captures the long-term utility of tokens. This is different from attention scores, which are query-dependent: they capture the short-term utility for predicting the next token and are recomputed at every step, making them local, myopic, and highly dependent on the transient decoding state.
 
-
 <a href="https://arxiv.org/pdf/2512.03324"><img src="https://img.shields.io/badge/arxiv-2512.03324-red?style=for-the-badge"></a>
 
+- **Official Code:** [GitHub - ngocbh/trimkv](https://github.com/ngocbh/trimkv)
+- **Paper:** [https://huggingface.co/papers/2605.09649](https://huggingface.co/papers/2605.09649)
 
 ### Why TRIM-KV?
 
@@ -62,8 +66,6 @@ And it's interpretable
 pip install -r requirements.txt
 ```
 
-This is a minimal set of requirements for training purposes. Additional dependencies may be needed for running specific experiments. We provided a full example of the environment used in our experiments in [`examples/env.yaml`](examples/env.yaml).
-
 ### Installation
 
 From the root of the repo:
@@ -72,7 +74,7 @@ From the root of the repo:
 git clone https://github.com/ngocbh/trimkv.git
 cd trimkv
 pip install -e .
-````
+```
 
 ---
 
@@ -84,7 +86,7 @@ from trimkv.models.qwen3 import TrimKVQwen3ForCausalLM
 from trimkv.cache_utils import TrimKVCache
 from transformers import AutoTokenizer
 
-model_path = "<TrimKV model_path here>"
+model_path = "ngocbh/TrimKV-Qwen3-14B-Math"
 download_from = "huggingface"  # options: "wandb", "local", "huggingface"
 
 model = TrimKVQwen3ForCausalLM.from_pretrained(
@@ -112,7 +114,7 @@ tokenizer = AutoTokenizer.from_pretrained(
 # Note: TRIM-KV uses TrimKVCache under the hood. So please pass TrimKVCache to model.generate
 ```
 
-For a runnable end-to-end example, see [`examples/test_qwen3.py`](examples/test_qwen3.py).
+For a runnable end-to-end example, see [`examples/test_qwen3.py`](https://github.com/ngocbh/trimkv/blob/main/examples/test_qwen3.py).
 
 ## Released Models
 
@@ -126,4 +128,13 @@ For a runnable end-to-end example, see [`examples/test_qwen3.py`](examples/test_
 | Phi-3-mini-128k-instruct | [TrimKV-Phi-3-mini-128k-instruct](https://huggingface.co/ngocbh/TrimKV-Phi-3-mini-128k-instruct) | LongAlpaca          |  128K  | 512 |
 | DeepSeek-R1-Distill-Llama-8B                    | [TrimKV-DeepSeek-R1-Distill-Llama-8B](https://huggingface.co/ngocbh/TrimKV-DeepSeek-R1-Distill-Llama-8B)           |  OpenR1-Math-220k         | 32K   | 256     |
 
----
+## Citation
+
+```bibtex
+@article{bui2025make,
+  title={Make Each Token Count: Towards Improving Long-Context Performance with KV Cache Eviction},
+  author={Bui, Ngoc and Nguyen, Hieu Trung and Cohan, Arman and Ying, Rex},
+  journal={arXiv preprint arXiv:2512.03324},
+  year={2025}
+}
+```