Spaces:
Runtime error
Runtime error
Update app.py
Browse files
app.py
CHANGED
|
@@ -44,6 +44,8 @@ from llama_index.core.node_parser.relational.base_element import (
|
|
| 44 |
)
|
| 45 |
from llama_index.core.schema import BaseNode, TextNode
|
| 46 |
|
|
|
|
|
|
|
| 47 |
# Implementations
|
| 48 |
|
| 49 |
# default_persist_directory = './chroma_HF/'
|
|
@@ -94,6 +96,7 @@ def initialize_llmchain(llm_model, temperature, max_tokens, top_k, vector_db, pr
|
|
| 94 |
if llm_model == "mistralai/Mixtral-8x7B-Instruct-v0.3":
|
| 95 |
llm = HuggingFaceEndpoint(
|
| 96 |
repo_id=llm_model,
|
|
|
|
| 97 |
temperature = temperature,
|
| 98 |
max_new_tokens = max_tokens,
|
| 99 |
top_k = top_k,
|
|
@@ -103,6 +106,7 @@ def initialize_llmchain(llm_model, temperature, max_tokens, top_k, vector_db, pr
|
|
| 103 |
raise gr.Error("LLM model is too large to be loaded automatically on free inference endpoint")
|
| 104 |
llm = HuggingFaceEndpoint(
|
| 105 |
repo_id=llm_model,
|
|
|
|
| 106 |
temperature = temperature,
|
| 107 |
max_new_tokens = max_tokens,
|
| 108 |
top_k = top_k,
|
|
@@ -110,6 +114,7 @@ def initialize_llmchain(llm_model, temperature, max_tokens, top_k, vector_db, pr
|
|
| 110 |
elif llm_model == "microsoft/phi-2":
|
| 111 |
llm = HuggingFaceEndpoint(
|
| 112 |
repo_id=llm_model,
|
|
|
|
| 113 |
temperature = temperature,
|
| 114 |
max_new_tokens = max_tokens,
|
| 115 |
top_k = top_k,
|
|
@@ -118,7 +123,8 @@ def initialize_llmchain(llm_model, temperature, max_tokens, top_k, vector_db, pr
|
|
| 118 |
)
|
| 119 |
elif llm_model == "TinyLlama/TinyLlama-1.1B-Chat-v1.0":
|
| 120 |
llm = HuggingFaceEndpoint(
|
| 121 |
-
repo_id=llm_model,
|
|
|
|
| 122 |
temperature = temperature,
|
| 123 |
max_new_tokens = 250,
|
| 124 |
top_k = top_k,
|
|
@@ -127,6 +133,7 @@ def initialize_llmchain(llm_model, temperature, max_tokens, top_k, vector_db, pr
|
|
| 127 |
raise gr.Error("Llama-2-7b-chat-hf model requires a Pro subscription...")
|
| 128 |
llm = HuggingFaceEndpoint(
|
| 129 |
repo_id=llm_model,
|
|
|
|
| 130 |
temperature = temperature,
|
| 131 |
max_new_tokens = max_tokens,
|
| 132 |
top_k = top_k,
|
|
@@ -134,6 +141,7 @@ def initialize_llmchain(llm_model, temperature, max_tokens, top_k, vector_db, pr
|
|
| 134 |
else:
|
| 135 |
llm = HuggingFaceEndpoint(
|
| 136 |
repo_id=llm_model,
|
|
|
|
| 137 |
temperature = temperature,
|
| 138 |
max_new_tokens = max_tokens,
|
| 139 |
top_k = top_k,
|
|
|
|
| 44 |
)
|
| 45 |
from llama_index.core.schema import BaseNode, TextNode
|
| 46 |
|
| 47 |
+
|
| 48 |
+
api_token = os.getenv("HF_TOKEN")
|
| 49 |
# Implementations
|
| 50 |
|
| 51 |
# default_persist_directory = './chroma_HF/'
|
|
|
|
| 96 |
if llm_model == "mistralai/Mixtral-8x7B-Instruct-v0.3":
|
| 97 |
llm = HuggingFaceEndpoint(
|
| 98 |
repo_id=llm_model,
|
| 99 |
+
huggingfacehub_api_token = api_token,
|
| 100 |
temperature = temperature,
|
| 101 |
max_new_tokens = max_tokens,
|
| 102 |
top_k = top_k,
|
|
|
|
| 106 |
raise gr.Error("LLM model is too large to be loaded automatically on free inference endpoint")
|
| 107 |
llm = HuggingFaceEndpoint(
|
| 108 |
repo_id=llm_model,
|
| 109 |
+
huggingfacehub_api_token = api_token,
|
| 110 |
temperature = temperature,
|
| 111 |
max_new_tokens = max_tokens,
|
| 112 |
top_k = top_k,
|
|
|
|
| 114 |
elif llm_model == "microsoft/phi-2":
|
| 115 |
llm = HuggingFaceEndpoint(
|
| 116 |
repo_id=llm_model,
|
| 117 |
+
huggingfacehub_api_token = api_token,
|
| 118 |
temperature = temperature,
|
| 119 |
max_new_tokens = max_tokens,
|
| 120 |
top_k = top_k,
|
|
|
|
| 123 |
)
|
| 124 |
elif llm_model == "TinyLlama/TinyLlama-1.1B-Chat-v1.0":
|
| 125 |
llm = HuggingFaceEndpoint(
|
| 126 |
+
repo_id=llm_model,
|
| 127 |
+
huggingfacehub_api_token = api_token,
|
| 128 |
temperature = temperature,
|
| 129 |
max_new_tokens = 250,
|
| 130 |
top_k = top_k,
|
|
|
|
| 133 |
raise gr.Error("Llama-2-7b-chat-hf model requires a Pro subscription...")
|
| 134 |
llm = HuggingFaceEndpoint(
|
| 135 |
repo_id=llm_model,
|
| 136 |
+
huggingfacehub_api_token = api_token,
|
| 137 |
temperature = temperature,
|
| 138 |
max_new_tokens = max_tokens,
|
| 139 |
top_k = top_k,
|
|
|
|
| 141 |
else:
|
| 142 |
llm = HuggingFaceEndpoint(
|
| 143 |
repo_id=llm_model,
|
| 144 |
+
huggingfacehub_api_token = api_token,
|
| 145 |
temperature = temperature,
|
| 146 |
max_new_tokens = max_tokens,
|
| 147 |
top_k = top_k,
|