Update README.md

README.md

@@ -15,9 +15,9 @@ tags:
 
 YingLong model is introduced in this [paper](xxxxxxxx) (coming soon). This version is pre-trained on **78B** time points. More details can be found at our [github](https://github.com/wxie9/YingLong/).
 
+
  
 ## Quickstart
-We currently only suppoert Ampere, Ada, or Hopper GPUs (e.g., A100, RTX 3090, RTX 4090, H100).
 
 ```bash
 pip install xformers transformers
@@ -26,13 +26,15 @@ git clone https://github.com/Dao-AILab/flash-attention && cd flash-attention
 cd csrc/rotary && pip install .
 cd ../layer_norm && pip install .
 ```
+The flash attention is not required. If you use V100 or other GPU doesn't support flash attention, just change the FlashAttention2Available = RequirementCache("flash-attn>=2.0.0.post1") to
+FlashAttention2Available = False in the model.py file. It should be able to run. 
 
 ```python
 import torch
 from transformers import AutoModelForCausalLM
 
 # load pretrain model
-model = AutoModelForCausalLM.from_pretrained('qcw1314/YingLong_50m', trust_remote_code=True,torch_dtype=torch.bfloat16).cuda()
+model = AutoModelForCausalLM.from_pretrained('qcw2333/YingLong_50m', trust_remote_code=True,torch_dtype=torch.bfloat16).cuda()
 
 # prepare input
 batch_size, lookback_length = 1, 2880
@@ -48,22 +50,8 @@ print(output.shape)
 A notebook example is also provided [here](https://github.com/wxie9/YingLong/blob/main/quickstart_zero_shot.ipynb). The sample codes for long-term forecasting tasks and gift-eval tasks are provided at [link](https://github.com/wxie9/YingLong/tree/main).
 
 
-## Specification
-
-<!-- * Architecture: Causal Transformer (Decoder-only)
-* Pre-training Scale: 260B time points
-* Context Length: up to 2880
-* Parameter Count: 84M
-* Patch Length: 96
-* Number of Layers: 8
- -->
-
-<!-- ## Acknowledgments
-
-This work was supported by the National Natural Science Foundation of China (62022050 and U2342217), the BNRist Innovation Fund (BNR2024RC01010), and the National Engineering Research Center for Big Data Software. 
+<!-- ## Specification -->
 
-The model is mostly built from the Internet public time series dataset, which comes from different research teams and providers. We sincerely thank all individuals and organizations who have contributed the data. Without their generous sharing, this model would not have existed.
- -->
 
 ## Citation
 Coming soon...