RAGchat

Build error

moriire commited on Apr 28, 2024

Commit

df24c3c

•

1 Parent(s): e06db31

Update app/llm.py

Files changed (1) hide show

app/llm.py CHANGED Viewed

@@ -32,7 +32,7 @@ class ChatModel(BaseModel):
 llm_chat = llama_cpp.Llama.from_pretrained(
     repo_id="moriire/healthcare-ai-q2_k",
     filename="*.gguf",
-    #tokenizer=llama_cpp.llama_tokenizer.LlamaHFTokenizer.from_pretrained("moriire/healthcare-ai-q4_k_m"),
     verbose=False,
      n_ctx=256,
      n_gpu_layers=0,
@@ -41,7 +41,7 @@ llm_chat = llama_cpp.Llama.from_pretrained(
 llm_generate = llama_cpp.Llama.from_pretrained(
     repo_id="moriire/healthcare-ai-q2_k",
     filename="*.gguf",
-    #tokenizer=llama_cpp.llama_tokenizer.LlamaHFTokenizer.from_pretrained("moriire/healthcare-ai-q8_0"),
     verbose=False,
      n_ctx=4096,
      n_gpu_layers=0,

 llm_chat = llama_cpp.Llama.from_pretrained(
     repo_id="moriire/healthcare-ai-q2_k",
     filename="*.gguf",
+    tokenizer=llama_cpp.llama_tokenizer.LlamaHFTokenizer.from_pretrained("moriire/healthcare-ai-q2_k"),
     verbose=False,
      n_ctx=256,
      n_gpu_layers=0,
 llm_generate = llama_cpp.Llama.from_pretrained(
     repo_id="moriire/healthcare-ai-q2_k",
     filename="*.gguf",
+    tokenizer=llama_cpp.llama_tokenizer.LlamaHFTokenizer.from_pretrained("moriire/healthcare-ai-q2_k"),
     verbose=False,
      n_ctx=4096,
      n_gpu_layers=0,