readme changes

README.md

@@ -1,7 +1,6 @@
 ---
 tags:
 - chemistry
-- medical
 widget:
 - text: <LIGAND>
   example_title: Generate molecule
@@ -21,8 +20,8 @@ The pretrained model is capable of zero-shot molecule generation and conformer g
 the distribution of the [Uni-Mol](https://github.com/deepmodeling/Uni-Mol) dataset.
 We also expose finetuned models:
 
-- For the model finetuned on GEOM-DRUGS, visit [huggingface.co/insilicomedicine/BindGPT-GEOM](https://huggingface.co/insilicomedicine/BindGPT-GEOM)
-- For the model finetuned with Reinforcement Learning on CrossDocked, visit [huggingface.co/insilicomedicine/BindGPT-RL](https://huggingface.co/insilicomedicine/BindGPT-RL)
+- For the model finetuned on GEOM-DRUGS, visit [huggingface.co/insilicomedicine/bindgpt_finetuned](https://huggingface.co/insilicomedicine/bindgpt_finetuned)
+- The model finetuned with Reinforcement Learning on CrossDocked is coming soon
 
 
 ## Unconditional generation
@@ -40,8 +39,8 @@ https://github.com/insilicomedicine/bindgpt
 # Download model from Hugginface:
 from transformers import AutoTokenizer, AutoModelForCausalLM
 
-tokenizer = AutoTokenizer.from_pretrained("artemZholus/BindGPT")
-model = AutoModelForCausalLM.from_pretrained("artemZholus/BindGPT").cuda()
+tokenizer = AutoTokenizer.from_pretrained("insilicomedicine/bindgpt_pretrained")
+model = AutoModelForCausalLM.from_pretrained("insilicomedicine/bindgpt_pretrained").cuda()
 
 # Generate 10 tokenized molecules without condition
 NUM_SAMPLES = 10