clementsan commited on
Commit
fa4eb33
1 Parent(s): 1ef8d7c

Raise errors for phi-2 and llama-2 models

Browse files
Files changed (1) hide show
  1. app.py +7 -0
app.py CHANGED
@@ -109,6 +109,7 @@ def initialize_llmchain(llm_model, temperature, max_tokens, top_k, vector_db, pr
109
  model_kwargs={"temperature": temperature, "max_new_tokens": max_tokens, "top_k": top_k, "load_in_8bit": True}
110
  )
111
  elif llm_model == "microsoft/phi-2":
 
112
  llm = HuggingFaceHub(
113
  repo_id=llm_model,
114
  model_kwargs={"temperature": temperature, "max_new_tokens": max_tokens, "top_k": top_k, "trust_remote_code": True, "torch_dtype": "auto"}
@@ -118,6 +119,12 @@ def initialize_llmchain(llm_model, temperature, max_tokens, top_k, vector_db, pr
118
  repo_id=llm_model,
119
  model_kwargs={"temperature": temperature, "max_new_tokens": 250, "top_k": top_k}
120
  )
 
 
 
 
 
 
121
  else:
122
  llm = HuggingFaceHub(
123
  repo_id=llm_model,
 
109
  model_kwargs={"temperature": temperature, "max_new_tokens": max_tokens, "top_k": top_k, "load_in_8bit": True}
110
  )
111
  elif llm_model == "microsoft/phi-2":
112
+ raise gr.Error("phi-2 model requires 'trust_remote_code=True', currently not supported by langchain HuggingFaceHub...")
113
  llm = HuggingFaceHub(
114
  repo_id=llm_model,
115
  model_kwargs={"temperature": temperature, "max_new_tokens": max_tokens, "top_k": top_k, "trust_remote_code": True, "torch_dtype": "auto"}
 
119
  repo_id=llm_model,
120
  model_kwargs={"temperature": temperature, "max_new_tokens": 250, "top_k": top_k}
121
  )
122
+ elif llm_model == "meta-llama/Llama-2-7b-chat-hf":
123
+ raise gr.Error("Llama-2-7b-chat-hf model requires a Pro subscription...")
124
+ llm = HuggingFaceHub(
125
+ repo_id=llm_model,
126
+ model_kwargs={"temperature": temperature, "max_new_tokens": max_tokens, "top_k": top_k}
127
+ )
128
  else:
129
  llm = HuggingFaceHub(
130
  repo_id=llm_model,