use hugging face inference api instead of loading pipeline since that wasn't working
Browse files
app.py
CHANGED
@@ -9,7 +9,7 @@ from langchain.chains import VectorDBQA
|
|
9 |
from huggingface_hub import snapshot_download
|
10 |
from langchain import OpenAI
|
11 |
from langchain import PromptTemplate
|
12 |
-
from langchain.llms import HuggingFacePipeline
|
13 |
|
14 |
|
15 |
BOOK_NAME = "1984"
|
@@ -84,11 +84,16 @@ def load_chain(model: Literal["openai", "GPT-NeoXT-Chat-Base-20B"] ="openai"):
|
|
84 |
llm = OpenAI(temperature=0.2)
|
85 |
|
86 |
if model=="GPT-NeoXT-Chat-Base-20B":
|
87 |
-
llm = HuggingFacePipeline.from_model_id(
|
88 |
-
|
|
|
|
|
|
|
|
|
|
|
89 |
task="text-generation",
|
90 |
model_kwargs={"temperature":0.2, "max_length":400}
|
91 |
-
)
|
92 |
|
93 |
# load chain
|
94 |
chain = VectorDBQA.from_chain_type(
|
@@ -153,6 +158,14 @@ with st.sidebar:
|
|
153 |
)
|
154 |
os.environ["OPENAI_API_KEY"] = api_key
|
155 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
156 |
st.markdown("---")
|
157 |
|
158 |
st.info("Based on [Talk2Book](https://github.com/batmanscode/Talk2Book)")
|
@@ -174,7 +187,7 @@ ask = col2.button("Ask")
|
|
174 |
|
175 |
if ask:
|
176 |
|
177 |
-
if
|
178 |
st.write(f"**{BOOK_NAME}:** Whoops looks like you forgot your API key buddy")
|
179 |
st.stop()
|
180 |
else:
|
|
|
9 |
from huggingface_hub import snapshot_download
|
10 |
from langchain import OpenAI
|
11 |
from langchain import PromptTemplate
|
12 |
+
from langchain.llms import HuggingFacePipeline, HuggingFaceHub
|
13 |
|
14 |
|
15 |
BOOK_NAME = "1984"
|
|
|
84 |
llm = OpenAI(temperature=0.2)
|
85 |
|
86 |
if model=="GPT-NeoXT-Chat-Base-20B":
|
87 |
+
# llm = HuggingFacePipeline.from_model_id(
|
88 |
+
# model_id="togethercomputer/GPT-NeoXT-Chat-Base-20B",
|
89 |
+
# task="text-generation",
|
90 |
+
# model_kwargs={"temperature":0.2, "max_length":400}
|
91 |
+
# )
|
92 |
+
llm = HuggingFaceHub(
|
93 |
+
repo_id="togethercomputer/GPT-NeoXT-Chat-Base-20B",
|
94 |
task="text-generation",
|
95 |
model_kwargs={"temperature":0.2, "max_length":400}
|
96 |
+
)
|
97 |
|
98 |
# load chain
|
99 |
chain = VectorDBQA.from_chain_type(
|
|
|
158 |
)
|
159 |
os.environ["OPENAI_API_KEY"] = api_key
|
160 |
|
161 |
+
|
162 |
+
if choice == "togethercomputer/GPT-NeoXT-Chat-Base-20B":
|
163 |
+
api_key = st.text_input(label = "Paste your Hugging Face Hub API key here to get started",
|
164 |
+
type = "password",
|
165 |
+
help = "This isn't saved 🙈"
|
166 |
+
)
|
167 |
+
os.environ["HUGGINGFACEHUB_API_TOKEN"] = api_key
|
168 |
+
|
169 |
st.markdown("---")
|
170 |
|
171 |
st.info("Based on [Talk2Book](https://github.com/batmanscode/Talk2Book)")
|
|
|
187 |
|
188 |
if ask:
|
189 |
|
190 |
+
if api_key is "":
|
191 |
st.write(f"**{BOOK_NAME}:** Whoops looks like you forgot your API key buddy")
|
192 |
st.stop()
|
193 |
else:
|