Spaces:
Sleeping
Sleeping
Tuchuanhuhuhu
commited on
Commit
·
03f9025
1
Parent(s):
1bfb00d
Added support for multi-modal Model: XMBot
Browse files- ChuanhuChatbot.py +3 -0
- modules/base_model.py +31 -18
- modules/models.py +110 -11
- modules/overwrites.py +55 -17
- modules/presets.py +4 -3
- modules/utils.py +10 -4
ChuanhuChatbot.py
CHANGED
@@ -12,6 +12,7 @@ from modules.presets import *
|
|
12 |
from modules.overwrites import *
|
13 |
from modules.models import get_model
|
14 |
|
|
|
15 |
gr.Chatbot.postprocess = postprocess
|
16 |
PromptHelper.compact_text_chunks = compact_text_chunks
|
17 |
|
@@ -321,6 +322,8 @@ with gr.Blocks(css=customCSS, theme=small_and_beautiful_theme) as demo:
|
|
321 |
submitBtn.click(**transfer_input_args).then(**chatgpt_predict_args).then(**end_outputing_args)
|
322 |
submitBtn.click(**get_usage_args)
|
323 |
|
|
|
|
|
324 |
emptyBtn.click(
|
325 |
reset,
|
326 |
inputs=[current_model],
|
|
|
12 |
from modules.overwrites import *
|
13 |
from modules.models import get_model
|
14 |
|
15 |
+
gr.Chatbot._postprocess_chat_messages = postprocess_chat_messages
|
16 |
gr.Chatbot.postprocess = postprocess
|
17 |
PromptHelper.compact_text_chunks = compact_text_chunks
|
18 |
|
|
|
322 |
submitBtn.click(**transfer_input_args).then(**chatgpt_predict_args).then(**end_outputing_args)
|
323 |
submitBtn.click(**get_usage_args)
|
324 |
|
325 |
+
index_files.change(handle_file_upload, [current_model, index_files, chatbot], [index_files, chatbot, status_display])
|
326 |
+
|
327 |
emptyBtn.click(
|
328 |
reset,
|
329 |
inputs=[current_model],
|
modules/base_model.py
CHANGED
@@ -8,6 +8,7 @@ import os
|
|
8 |
import sys
|
9 |
import requests
|
10 |
import urllib3
|
|
|
11 |
|
12 |
from tqdm import tqdm
|
13 |
import colorama
|
@@ -28,6 +29,7 @@ class ModelType(Enum):
|
|
28 |
OpenAI = 0
|
29 |
ChatGLM = 1
|
30 |
LLaMA = 2
|
|
|
31 |
|
32 |
@classmethod
|
33 |
def get_type(cls, model_name: str):
|
@@ -39,6 +41,8 @@ class ModelType(Enum):
|
|
39 |
model_type = ModelType.ChatGLM
|
40 |
elif "llama" in model_name_lower or "alpaca" in model_name_lower:
|
41 |
model_type = ModelType.LLaMA
|
|
|
|
|
42 |
else:
|
43 |
model_type = ModelType.Unknown
|
44 |
return model_type
|
@@ -164,10 +168,19 @@ class BaseLLMModel:
|
|
164 |
status_text = self.token_message()
|
165 |
return chatbot, status_text
|
166 |
|
167 |
-
def
|
168 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
169 |
display_append = []
|
170 |
limited_context = False
|
|
|
171 |
if files:
|
172 |
from llama_index.indices.vector_store.base_query import GPTVectorStoreIndexQuery
|
173 |
from llama_index.indices.query.schema import QueryBundle
|
@@ -180,12 +193,11 @@ class BaseLLMModel:
|
|
180 |
OpenAIEmbedding,
|
181 |
)
|
182 |
limited_context = True
|
183 |
-
|
184 |
-
msg = "加载索引中……(这可能需要几分钟)"
|
185 |
logging.info(msg)
|
186 |
# yield chatbot + [(inputs, "")], msg
|
187 |
index = construct_index(self.api_key, file_src=files)
|
188 |
-
assert index is not None, "
|
189 |
msg = "索引获取成功,生成回答中……"
|
190 |
logging.info(msg)
|
191 |
if local_embedding or self.model_type != ModelType.OpenAI:
|
@@ -212,22 +224,21 @@ class BaseLLMModel:
|
|
212 |
vector_store=index._vector_store,
|
213 |
docstore=index._docstore,
|
214 |
)
|
215 |
-
query_bundle = QueryBundle(
|
216 |
nodes = query_object.retrieve(query_bundle)
|
217 |
reference_results = [n.node.text for n in nodes]
|
218 |
reference_results = add_source_numbers(reference_results, use_source=False)
|
219 |
display_append = add_details(reference_results)
|
220 |
display_append = "\n\n" + "".join(display_append)
|
221 |
-
|
222 |
replace_today(PROMPT_TEMPLATE)
|
223 |
-
.replace("{query_str}",
|
224 |
.replace("{context_str}", "\n\n".join(reference_results))
|
225 |
.replace("{reply_language}", reply_language)
|
226 |
)
|
227 |
elif use_websearch:
|
228 |
limited_context = True
|
229 |
-
search_results = ddg(
|
230 |
-
old_inputs = inputs
|
231 |
reference_results = []
|
232 |
for idx, result in enumerate(search_results):
|
233 |
logging.debug(f"搜索结果{idx + 1}:{result}")
|
@@ -238,15 +249,15 @@ class BaseLLMModel:
|
|
238 |
)
|
239 |
reference_results = add_source_numbers(reference_results)
|
240 |
display_append = "\n\n" + "".join(display_append)
|
241 |
-
|
242 |
replace_today(WEBSEARCH_PTOMPT_TEMPLATE)
|
243 |
-
.replace("{query}",
|
244 |
.replace("{web_results}", "\n\n".join(reference_results))
|
245 |
.replace("{reply_language}", reply_language)
|
246 |
)
|
247 |
else:
|
248 |
display_append = ""
|
249 |
-
return limited_context,
|
250 |
|
251 |
def predict(
|
252 |
self,
|
@@ -259,16 +270,17 @@ class BaseLLMModel:
|
|
259 |
should_check_token_count=True,
|
260 |
): # repetition_penalty, top_k
|
261 |
|
262 |
-
|
263 |
logging.info(
|
264 |
"输入为:" + colorama.Fore.BLUE + f"{inputs}" + colorama.Style.RESET_ALL
|
265 |
)
|
266 |
if should_check_token_count:
|
267 |
-
yield chatbot + [(inputs, "")],
|
268 |
if reply_language == "跟随问题语言(不稳定)":
|
269 |
reply_language = "the same language as the question, such as English, 中文, 日本語, Español, Français, or Deutsch."
|
270 |
|
271 |
-
limited_context,
|
|
|
272 |
|
273 |
if (
|
274 |
self.need_api_key and
|
@@ -303,7 +315,7 @@ class BaseLLMModel:
|
|
303 |
iter = self.stream_next_chatbot(
|
304 |
inputs,
|
305 |
chatbot,
|
306 |
-
fake_input=
|
307 |
display_append=display_append,
|
308 |
)
|
309 |
for chatbot, status_text in iter:
|
@@ -313,11 +325,12 @@ class BaseLLMModel:
|
|
313 |
chatbot, status_text = self.next_chatbot_at_once(
|
314 |
inputs,
|
315 |
chatbot,
|
316 |
-
fake_input=
|
317 |
display_append=display_append,
|
318 |
)
|
319 |
yield chatbot, status_text
|
320 |
except Exception as e:
|
|
|
321 |
status_text = STANDARD_ERROR_MSG + str(e)
|
322 |
yield chatbot, status_text
|
323 |
|
|
|
8 |
import sys
|
9 |
import requests
|
10 |
import urllib3
|
11 |
+
import traceback
|
12 |
|
13 |
from tqdm import tqdm
|
14 |
import colorama
|
|
|
29 |
OpenAI = 0
|
30 |
ChatGLM = 1
|
31 |
LLaMA = 2
|
32 |
+
XMBot = 3
|
33 |
|
34 |
@classmethod
|
35 |
def get_type(cls, model_name: str):
|
|
|
41 |
model_type = ModelType.ChatGLM
|
42 |
elif "llama" in model_name_lower or "alpaca" in model_name_lower:
|
43 |
model_type = ModelType.LLaMA
|
44 |
+
elif "xmbot" in model_name_lower:
|
45 |
+
model_type = ModelType.XMBot
|
46 |
else:
|
47 |
model_type = ModelType.Unknown
|
48 |
return model_type
|
|
|
168 |
status_text = self.token_message()
|
169 |
return chatbot, status_text
|
170 |
|
171 |
+
def handle_file_upload(self, files, chatbot):
|
172 |
+
"""if the model accepts multi modal input, implement this function"""
|
173 |
+
status = gr.Markdown.update()
|
174 |
+
if files:
|
175 |
+
construct_index(self.api_key, file_src=files)
|
176 |
+
status = "索引构建完成"
|
177 |
+
return gr.Files.update(), chatbot, status
|
178 |
+
|
179 |
+
def prepare_inputs(self, real_inputs, use_websearch, files, reply_language, chatbot):
|
180 |
+
fake_inputs = None
|
181 |
display_append = []
|
182 |
limited_context = False
|
183 |
+
fake_inputs = real_inputs
|
184 |
if files:
|
185 |
from llama_index.indices.vector_store.base_query import GPTVectorStoreIndexQuery
|
186 |
from llama_index.indices.query.schema import QueryBundle
|
|
|
193 |
OpenAIEmbedding,
|
194 |
)
|
195 |
limited_context = True
|
196 |
+
msg = "加载索引中……"
|
|
|
197 |
logging.info(msg)
|
198 |
# yield chatbot + [(inputs, "")], msg
|
199 |
index = construct_index(self.api_key, file_src=files)
|
200 |
+
assert index is not None, "获取索引失败"
|
201 |
msg = "索引获取成功,生成回答中……"
|
202 |
logging.info(msg)
|
203 |
if local_embedding or self.model_type != ModelType.OpenAI:
|
|
|
224 |
vector_store=index._vector_store,
|
225 |
docstore=index._docstore,
|
226 |
)
|
227 |
+
query_bundle = QueryBundle(real_inputs)
|
228 |
nodes = query_object.retrieve(query_bundle)
|
229 |
reference_results = [n.node.text for n in nodes]
|
230 |
reference_results = add_source_numbers(reference_results, use_source=False)
|
231 |
display_append = add_details(reference_results)
|
232 |
display_append = "\n\n" + "".join(display_append)
|
233 |
+
real_inputs = (
|
234 |
replace_today(PROMPT_TEMPLATE)
|
235 |
+
.replace("{query_str}", real_inputs)
|
236 |
.replace("{context_str}", "\n\n".join(reference_results))
|
237 |
.replace("{reply_language}", reply_language)
|
238 |
)
|
239 |
elif use_websearch:
|
240 |
limited_context = True
|
241 |
+
search_results = ddg(real_inputs, max_results=5)
|
|
|
242 |
reference_results = []
|
243 |
for idx, result in enumerate(search_results):
|
244 |
logging.debug(f"搜索结果{idx + 1}:{result}")
|
|
|
249 |
)
|
250 |
reference_results = add_source_numbers(reference_results)
|
251 |
display_append = "\n\n" + "".join(display_append)
|
252 |
+
real_inputs = (
|
253 |
replace_today(WEBSEARCH_PTOMPT_TEMPLATE)
|
254 |
+
.replace("{query}", real_inputs)
|
255 |
.replace("{web_results}", "\n\n".join(reference_results))
|
256 |
.replace("{reply_language}", reply_language)
|
257 |
)
|
258 |
else:
|
259 |
display_append = ""
|
260 |
+
return limited_context, fake_inputs, display_append, real_inputs, chatbot
|
261 |
|
262 |
def predict(
|
263 |
self,
|
|
|
270 |
should_check_token_count=True,
|
271 |
): # repetition_penalty, top_k
|
272 |
|
273 |
+
status_text = "开始生成回答……"
|
274 |
logging.info(
|
275 |
"输入为:" + colorama.Fore.BLUE + f"{inputs}" + colorama.Style.RESET_ALL
|
276 |
)
|
277 |
if should_check_token_count:
|
278 |
+
yield chatbot + [(inputs, "")], status_text
|
279 |
if reply_language == "跟随问题语言(不稳定)":
|
280 |
reply_language = "the same language as the question, such as English, 中文, 日本語, Español, Français, or Deutsch."
|
281 |
|
282 |
+
limited_context, fake_inputs, display_append, inputs, chatbot = self.prepare_inputs(real_inputs=inputs, use_websearch=use_websearch, files=files, reply_language=reply_language, chatbot=chatbot)
|
283 |
+
yield chatbot + [(fake_inputs, "")], status_text
|
284 |
|
285 |
if (
|
286 |
self.need_api_key and
|
|
|
315 |
iter = self.stream_next_chatbot(
|
316 |
inputs,
|
317 |
chatbot,
|
318 |
+
fake_input=fake_inputs,
|
319 |
display_append=display_append,
|
320 |
)
|
321 |
for chatbot, status_text in iter:
|
|
|
325 |
chatbot, status_text = self.next_chatbot_at_once(
|
326 |
inputs,
|
327 |
chatbot,
|
328 |
+
fake_input=fake_inputs,
|
329 |
display_append=display_append,
|
330 |
)
|
331 |
yield chatbot, status_text
|
332 |
except Exception as e:
|
333 |
+
traceback.print_exc()
|
334 |
status_text = STANDARD_ERROR_MSG + str(e)
|
335 |
yield chatbot, status_text
|
336 |
|
modules/models.py
CHANGED
@@ -16,6 +16,7 @@ from duckduckgo_search import ddg
|
|
16 |
import asyncio
|
17 |
import aiohttp
|
18 |
from enum import Enum
|
|
|
19 |
|
20 |
from .presets import *
|
21 |
from .llama_func import *
|
@@ -75,7 +76,8 @@ class OpenAIClient(BaseLLMModel):
|
|
75 |
def billing_info(self):
|
76 |
try:
|
77 |
curr_time = datetime.datetime.now()
|
78 |
-
last_day_of_month = get_last_day_of_month(
|
|
|
79 |
first_day_of_month = curr_time.replace(day=1).strftime("%Y-%m-%d")
|
80 |
usage_url = f"{shared.state.usage_api_url}?start_date={first_day_of_month}&end_date={last_day_of_month}"
|
81 |
try:
|
@@ -112,7 +114,8 @@ class OpenAIClient(BaseLLMModel):
|
|
112 |
openai_api_key = self.api_key
|
113 |
system_prompt = self.system_prompt
|
114 |
history = self.history
|
115 |
-
logging.debug(colorama.Fore.YELLOW +
|
|
|
116 |
headers = {
|
117 |
"Content-Type": "application/json",
|
118 |
"Authorization": f"Bearer {openai_api_key}",
|
@@ -217,7 +220,7 @@ class ChatGLM_Client(BaseLLMModel):
|
|
217 |
global CHATGLM_TOKENIZER, CHATGLM_MODEL
|
218 |
if CHATGLM_TOKENIZER is None or CHATGLM_MODEL is None:
|
219 |
system_name = platform.system()
|
220 |
-
model_path=None
|
221 |
if os.path.exists("models"):
|
222 |
model_dirs = os.listdir("models")
|
223 |
if model_name in model_dirs:
|
@@ -257,16 +260,19 @@ class ChatGLM_Client(BaseLLMModel):
|
|
257 |
def _get_glm_style_input(self):
|
258 |
history = [x["content"] for x in self.history]
|
259 |
query = history.pop()
|
260 |
-
logging.debug(colorama.Fore.YELLOW +
|
|
|
261 |
assert (
|
262 |
len(history) % 2 == 0
|
263 |
), f"History should be even length. current history is: {history}"
|
264 |
-
history = [[history[i], history[i + 1]]
|
|
|
265 |
return history, query
|
266 |
|
267 |
def get_answer_at_once(self):
|
268 |
history, query = self._get_glm_style_input()
|
269 |
-
response, _ = CHATGLM_MODEL.chat(
|
|
|
270 |
return response, len(response)
|
271 |
|
272 |
def get_answer_stream_iter(self):
|
@@ -315,8 +321,10 @@ class LLaMA_Client(BaseLLMModel):
|
|
315 |
# raise Exception(f"models目录下没有这个模型: {model_name}")
|
316 |
if lora_path is not None:
|
317 |
lora_path = f"lora/{lora_path}"
|
318 |
-
model_args = ModelArguments(model_name_or_path=model_source, lora_model_path=lora_path, model_type=None, config_overrides=None, config_name=None, tokenizer_name=None, cache_dir=None,
|
319 |
-
|
|
|
|
|
320 |
|
321 |
with open(pipeline_args.deepspeed, "r") as f:
|
322 |
ds_config = json.load(f)
|
@@ -341,7 +349,6 @@ class LLaMA_Client(BaseLLMModel):
|
|
341 |
# " unconditionally."
|
342 |
# )
|
343 |
|
344 |
-
|
345 |
def _get_llama_style_input(self):
|
346 |
history = []
|
347 |
instruction = ""
|
@@ -379,7 +386,8 @@ class LLaMA_Client(BaseLLMModel):
|
|
379 |
step = 1
|
380 |
for _ in range(0, self.max_generation_token, step):
|
381 |
input_dataset = self.dataset.from_dict(
|
382 |
-
{"type": "text_only", "instances": [
|
|
|
383 |
)
|
384 |
output_dataset = LLAMA_INFERENCER.inference(
|
385 |
model=LLAMA_MODEL,
|
@@ -394,6 +402,94 @@ class LLaMA_Client(BaseLLMModel):
|
|
394 |
yield partial_text
|
395 |
|
396 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
397 |
def get_model(
|
398 |
model_name,
|
399 |
lora_model_path=None,
|
@@ -429,7 +525,8 @@ def get_model(
|
|
429 |
logging.info(msg)
|
430 |
lora_selector_visibility = True
|
431 |
if os.path.isdir("lora"):
|
432 |
-
lora_choices = get_file_names(
|
|
|
433 |
lora_choices = ["No LoRA"] + lora_choices
|
434 |
elif model_type == ModelType.LLaMA and lora_model_path != "":
|
435 |
logging.info(f"正在加载LLaMA模型: {model_name} + {lora_model_path}")
|
@@ -440,6 +537,8 @@ def get_model(
|
|
440 |
else:
|
441 |
msg += f" + {lora_model_path}"
|
442 |
model = LLaMA_Client(model_name, lora_model_path)
|
|
|
|
|
443 |
elif model_type == ModelType.Unknown:
|
444 |
raise ValueError(f"未知模型: {model_name}")
|
445 |
logging.info(msg)
|
|
|
16 |
import asyncio
|
17 |
import aiohttp
|
18 |
from enum import Enum
|
19 |
+
import uuid
|
20 |
|
21 |
from .presets import *
|
22 |
from .llama_func import *
|
|
|
76 |
def billing_info(self):
|
77 |
try:
|
78 |
curr_time = datetime.datetime.now()
|
79 |
+
last_day_of_month = get_last_day_of_month(
|
80 |
+
curr_time).strftime("%Y-%m-%d")
|
81 |
first_day_of_month = curr_time.replace(day=1).strftime("%Y-%m-%d")
|
82 |
usage_url = f"{shared.state.usage_api_url}?start_date={first_day_of_month}&end_date={last_day_of_month}"
|
83 |
try:
|
|
|
114 |
openai_api_key = self.api_key
|
115 |
system_prompt = self.system_prompt
|
116 |
history = self.history
|
117 |
+
logging.debug(colorama.Fore.YELLOW +
|
118 |
+
f"{history}" + colorama.Fore.RESET)
|
119 |
headers = {
|
120 |
"Content-Type": "application/json",
|
121 |
"Authorization": f"Bearer {openai_api_key}",
|
|
|
220 |
global CHATGLM_TOKENIZER, CHATGLM_MODEL
|
221 |
if CHATGLM_TOKENIZER is None or CHATGLM_MODEL is None:
|
222 |
system_name = platform.system()
|
223 |
+
model_path = None
|
224 |
if os.path.exists("models"):
|
225 |
model_dirs = os.listdir("models")
|
226 |
if model_name in model_dirs:
|
|
|
260 |
def _get_glm_style_input(self):
|
261 |
history = [x["content"] for x in self.history]
|
262 |
query = history.pop()
|
263 |
+
logging.debug(colorama.Fore.YELLOW +
|
264 |
+
f"{history}" + colorama.Fore.RESET)
|
265 |
assert (
|
266 |
len(history) % 2 == 0
|
267 |
), f"History should be even length. current history is: {history}"
|
268 |
+
history = [[history[i], history[i + 1]]
|
269 |
+
for i in range(0, len(history), 2)]
|
270 |
return history, query
|
271 |
|
272 |
def get_answer_at_once(self):
|
273 |
history, query = self._get_glm_style_input()
|
274 |
+
response, _ = CHATGLM_MODEL.chat(
|
275 |
+
CHATGLM_TOKENIZER, query, history=history)
|
276 |
return response, len(response)
|
277 |
|
278 |
def get_answer_stream_iter(self):
|
|
|
321 |
# raise Exception(f"models目录下没有这个模型: {model_name}")
|
322 |
if lora_path is not None:
|
323 |
lora_path = f"lora/{lora_path}"
|
324 |
+
model_args = ModelArguments(model_name_or_path=model_source, lora_model_path=lora_path, model_type=None, config_overrides=None, config_name=None, tokenizer_name=None, cache_dir=None,
|
325 |
+
use_fast_tokenizer=True, model_revision='main', use_auth_token=False, torch_dtype=None, use_lora=False, lora_r=8, lora_alpha=32, lora_dropout=0.1, use_ram_optimized_load=True)
|
326 |
+
pipeline_args = InferencerArguments(
|
327 |
+
local_rank=0, random_seed=1, deepspeed='configs/ds_config_chatbot.json', mixed_precision='bf16')
|
328 |
|
329 |
with open(pipeline_args.deepspeed, "r") as f:
|
330 |
ds_config = json.load(f)
|
|
|
349 |
# " unconditionally."
|
350 |
# )
|
351 |
|
|
|
352 |
def _get_llama_style_input(self):
|
353 |
history = []
|
354 |
instruction = ""
|
|
|
386 |
step = 1
|
387 |
for _ in range(0, self.max_generation_token, step):
|
388 |
input_dataset = self.dataset.from_dict(
|
389 |
+
{"type": "text_only", "instances": [
|
390 |
+
{"text": context + partial_text}]}
|
391 |
)
|
392 |
output_dataset = LLAMA_INFERENCER.inference(
|
393 |
model=LLAMA_MODEL,
|
|
|
402 |
yield partial_text
|
403 |
|
404 |
|
405 |
+
class XMBot_Client(BaseLLMModel):
|
406 |
+
def __init__(self, api_key):
|
407 |
+
super().__init__(model_name="xmbot")
|
408 |
+
self.api_key = api_key
|
409 |
+
self.session_id = None
|
410 |
+
self.reset()
|
411 |
+
self.image_bytes = None
|
412 |
+
self.image_path = None
|
413 |
+
self.xm_history = []
|
414 |
+
self.url = "https://xmbot.net/web"
|
415 |
+
|
416 |
+
def reset(self):
|
417 |
+
self.session_id = str(uuid.uuid4())
|
418 |
+
return [], "已重置"
|
419 |
+
|
420 |
+
def try_read_image(self, filepath):
|
421 |
+
import base64
|
422 |
+
|
423 |
+
def is_image_file(filepath):
|
424 |
+
# 判断文件是否为图片
|
425 |
+
valid_image_extensions = [".jpg", ".jpeg", ".png", ".bmp", ".gif", ".tiff"]
|
426 |
+
file_extension = os.path.splitext(filepath)[1].lower()
|
427 |
+
return file_extension in valid_image_extensions
|
428 |
+
|
429 |
+
def read_image_as_bytes(filepath):
|
430 |
+
# 读取图片文件并返回比特流
|
431 |
+
with open(filepath, "rb") as f:
|
432 |
+
image_bytes = f.read()
|
433 |
+
return image_bytes
|
434 |
+
|
435 |
+
if is_image_file(filepath):
|
436 |
+
logging.info(f"读取图片文件: {filepath}")
|
437 |
+
image_bytes = read_image_as_bytes(filepath)
|
438 |
+
base64_encoded_image = base64.b64encode(image_bytes).decode()
|
439 |
+
self.image_bytes = base64_encoded_image
|
440 |
+
self.image_path = filepath
|
441 |
+
else:
|
442 |
+
self.image_bytes = None
|
443 |
+
self.image_path = None
|
444 |
+
|
445 |
+
def prepare_inputs(self, real_inputs, use_websearch, files, reply_language, chatbot):
|
446 |
+
fake_inputs = real_inputs
|
447 |
+
display_append = ""
|
448 |
+
limited_context = False
|
449 |
+
return limited_context, fake_inputs, display_append, real_inputs, chatbot
|
450 |
+
|
451 |
+
def handle_file_upload(self, files, chatbot):
|
452 |
+
"""if the model accepts multi modal input, implement this function"""
|
453 |
+
if files:
|
454 |
+
for file in files:
|
455 |
+
if file.name:
|
456 |
+
logging.info(f"尝试读取图像: {file.name}")
|
457 |
+
self.try_read_image(file.name)
|
458 |
+
if self.image_path is not None:
|
459 |
+
chatbot = chatbot + [((self.image_path,), None)]
|
460 |
+
if self.image_bytes is not None:
|
461 |
+
logging.info("使用图片作为输入")
|
462 |
+
conv_id = str(uuid.uuid4())
|
463 |
+
data = {
|
464 |
+
"user_id": self.api_key,
|
465 |
+
"session_id": self.session_id,
|
466 |
+
"uuid": conv_id,
|
467 |
+
"data_type": "imgbase64",
|
468 |
+
"data": self.image_bytes
|
469 |
+
}
|
470 |
+
# response = requests.post(self.url, json=data)
|
471 |
+
# response = json.loads(response.text)
|
472 |
+
# logging.info(f"图片回复: {response['data']}")
|
473 |
+
logging.info("发送了图片")
|
474 |
+
return None, chatbot, None
|
475 |
+
|
476 |
+
def get_answer_at_once(self):
|
477 |
+
question = self.history[-1]["content"]
|
478 |
+
conv_id = str(uuid.uuid4())
|
479 |
+
data = {
|
480 |
+
"user_id": self.api_key,
|
481 |
+
"session_id": self.session_id,
|
482 |
+
"uuid": conv_id,
|
483 |
+
"data_type": "text",
|
484 |
+
"data": question
|
485 |
+
}
|
486 |
+
response = requests.post(self.url, json=data)
|
487 |
+
response = json.loads(response.text)
|
488 |
+
return response["data"], len(response["data"])
|
489 |
+
|
490 |
+
|
491 |
+
|
492 |
+
|
493 |
def get_model(
|
494 |
model_name,
|
495 |
lora_model_path=None,
|
|
|
525 |
logging.info(msg)
|
526 |
lora_selector_visibility = True
|
527 |
if os.path.isdir("lora"):
|
528 |
+
lora_choices = get_file_names(
|
529 |
+
"lora", plain=True, filetypes=[""])
|
530 |
lora_choices = ["No LoRA"] + lora_choices
|
531 |
elif model_type == ModelType.LLaMA and lora_model_path != "":
|
532 |
logging.info(f"正在加载LLaMA模型: {model_name} + {lora_model_path}")
|
|
|
537 |
else:
|
538 |
msg += f" + {lora_model_path}"
|
539 |
model = LLaMA_Client(model_name, lora_model_path)
|
540 |
+
elif model_type == ModelType.XMBot:
|
541 |
+
model = XMBot_Client(api_key=access_key)
|
542 |
elif model_type == ModelType.Unknown:
|
543 |
raise ValueError(f"未知模型: {model_name}")
|
544 |
logging.info(msg)
|
modules/overwrites.py
CHANGED
@@ -4,6 +4,7 @@ import logging
|
|
4 |
from llama_index import Prompt
|
5 |
from typing import List, Tuple
|
6 |
import mdtex2html
|
|
|
7 |
|
8 |
from modules.presets import *
|
9 |
from modules.llama_func import *
|
@@ -20,23 +21,60 @@ def compact_text_chunks(self, prompt: Prompt, text_chunks: List[str]) -> List[st
|
|
20 |
|
21 |
|
22 |
def postprocess(
|
23 |
-
|
24 |
-
|
25 |
-
|
26 |
-
|
27 |
-
|
28 |
-
|
29 |
-
|
30 |
-
|
31 |
-
|
32 |
-
|
33 |
-
|
34 |
-
|
35 |
-
|
36 |
-
|
37 |
-
|
38 |
-
|
39 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
40 |
|
41 |
with open("./assets/custom.js", "r", encoding="utf-8") as f, open("./assets/Kelpy-Codos.js", "r", encoding="utf-8") as f2:
|
42 |
customJS = f.read()
|
|
|
4 |
from llama_index import Prompt
|
5 |
from typing import List, Tuple
|
6 |
import mdtex2html
|
7 |
+
from gradio_client import utils as client_utils
|
8 |
|
9 |
from modules.presets import *
|
10 |
from modules.llama_func import *
|
|
|
21 |
|
22 |
|
23 |
def postprocess(
|
24 |
+
self,
|
25 |
+
y: List[List[str | Tuple[str] | Tuple[str, str] | None] | Tuple],
|
26 |
+
) -> List[List[str | Dict | None]]:
|
27 |
+
"""
|
28 |
+
Parameters:
|
29 |
+
y: List of lists representing the message and response pairs. Each message and response should be a string, which may be in Markdown format. It can also be a tuple whose first element is a string filepath or URL to an image/video/audio, and second (optional) element is the alt text, in which case the media file is displayed. It can also be None, in which case that message is not displayed.
|
30 |
+
Returns:
|
31 |
+
List of lists representing the message and response. Each message and response will be a string of HTML, or a dictionary with media information. Or None if the message is not to be displayed.
|
32 |
+
"""
|
33 |
+
if y is None:
|
34 |
+
return []
|
35 |
+
processed_messages = []
|
36 |
+
for message_pair in y:
|
37 |
+
assert isinstance(
|
38 |
+
message_pair, (tuple, list)
|
39 |
+
), f"Expected a list of lists or list of tuples. Received: {message_pair}"
|
40 |
+
assert (
|
41 |
+
len(message_pair) == 2
|
42 |
+
), f"Expected a list of lists of length 2 or list of tuples of length 2. Received: {message_pair}"
|
43 |
+
|
44 |
+
processed_messages.append(
|
45 |
+
[
|
46 |
+
self._postprocess_chat_messages(message_pair[0], "user"),
|
47 |
+
self._postprocess_chat_messages(message_pair[1], "bot"),
|
48 |
+
]
|
49 |
+
)
|
50 |
+
return processed_messages
|
51 |
+
|
52 |
+
def postprocess_chat_messages(
|
53 |
+
self, chat_message: str | Tuple | List | None, message_type: str
|
54 |
+
) -> str | Dict | None:
|
55 |
+
if chat_message is None:
|
56 |
+
return None
|
57 |
+
elif isinstance(chat_message, (tuple, list)):
|
58 |
+
filepath = chat_message[0]
|
59 |
+
mime_type = client_utils.get_mimetype(filepath)
|
60 |
+
filepath = self.make_temp_copy_if_needed(filepath)
|
61 |
+
return {
|
62 |
+
"name": filepath,
|
63 |
+
"mime_type": mime_type,
|
64 |
+
"alt_text": chat_message[1] if len(chat_message) > 1 else None,
|
65 |
+
"data": None, # These last two fields are filled in by the frontend
|
66 |
+
"is_file": True,
|
67 |
+
}
|
68 |
+
elif isinstance(chat_message, str):
|
69 |
+
if message_type == "bot":
|
70 |
+
if not detect_converted_mark(chat_message):
|
71 |
+
chat_message = convert_mdtext(chat_message)
|
72 |
+
elif message_type == "user":
|
73 |
+
if not detect_converted_mark(chat_message):
|
74 |
+
chat_message = convert_asis(chat_message)
|
75 |
+
return chat_message
|
76 |
+
else:
|
77 |
+
raise ValueError(f"Invalid message for Chatbot component: {chat_message}")
|
78 |
|
79 |
with open("./assets/custom.js", "r", encoding="utf-8") as f, open("./assets/Kelpy-Codos.js", "r", encoding="utf-8") as f2:
|
80 |
customJS = f.read()
|
modules/presets.py
CHANGED
@@ -29,7 +29,7 @@ PROXY_ERROR_MSG = "代理错误,无法获取对话。" # 代理错误
|
|
29 |
SSL_ERROR_PROMPT = "SSL错误,无法获取对话。" # SSL 错误
|
30 |
NO_APIKEY_MSG = "API key为空,请检查是否输入正确。" # API key 长度不足 51 位
|
31 |
NO_INPUT_MSG = "请输入对话内容。" # 未输入对话内容
|
32 |
-
BILLING_NOT_APPLICABLE_MSG = "
|
33 |
|
34 |
TIMEOUT_STREAMING = 60 # 流式对话时的超时时间
|
35 |
TIMEOUT_ALL = 200 # 非流式对话时的超时时间
|
@@ -72,6 +72,7 @@ MODELS = [
|
|
72 |
"gpt-4-0314",
|
73 |
"gpt-4-32k",
|
74 |
"gpt-4-32k-0314",
|
|
|
75 |
"chatglm-6b",
|
76 |
"chatglm-6b-int4",
|
77 |
"chatglm-6b-int4-qe",
|
@@ -85,6 +86,8 @@ MODELS = [
|
|
85 |
"llama-65b-hf",
|
86 |
] # 可选的模型
|
87 |
|
|
|
|
|
88 |
os.makedirs("models", exist_ok=True)
|
89 |
os.makedirs("lora", exist_ok=True)
|
90 |
os.makedirs("history", exist_ok=True)
|
@@ -93,8 +96,6 @@ for dir_name in os.listdir("models"):
|
|
93 |
if dir_name not in MODELS:
|
94 |
MODELS.append(dir_name)
|
95 |
|
96 |
-
DEFAULT_MODEL = 0 # 默认的模型在MODELS中的序号,从0开始数
|
97 |
-
|
98 |
MODEL_TOKEN_LIMIT = {
|
99 |
"gpt-3.5-turbo": 4096,
|
100 |
"gpt-3.5-turbo-0301": 4096,
|
|
|
29 |
SSL_ERROR_PROMPT = "SSL错误,无法获取对话。" # SSL 错误
|
30 |
NO_APIKEY_MSG = "API key为空,请检查是否输入正确。" # API key 长度不足 51 位
|
31 |
NO_INPUT_MSG = "请输入对话内容。" # 未输入对话内容
|
32 |
+
BILLING_NOT_APPLICABLE_MSG = "账单信息不适用" # 本地运行的模型返回的账单信息
|
33 |
|
34 |
TIMEOUT_STREAMING = 60 # 流式对话时的超时时间
|
35 |
TIMEOUT_ALL = 200 # 非流式对话时的超时时间
|
|
|
72 |
"gpt-4-0314",
|
73 |
"gpt-4-32k",
|
74 |
"gpt-4-32k-0314",
|
75 |
+
"xmbot",
|
76 |
"chatglm-6b",
|
77 |
"chatglm-6b-int4",
|
78 |
"chatglm-6b-int4-qe",
|
|
|
86 |
"llama-65b-hf",
|
87 |
] # 可选的模型
|
88 |
|
89 |
+
DEFAULT_MODEL = 0 # 默认的模型在MODELS中的序号,从0开始数
|
90 |
+
|
91 |
os.makedirs("models", exist_ok=True)
|
92 |
os.makedirs("lora", exist_ok=True)
|
93 |
os.makedirs("history", exist_ok=True)
|
|
|
96 |
if dir_name not in MODELS:
|
97 |
MODELS.append(dir_name)
|
98 |
|
|
|
|
|
99 |
MODEL_TOKEN_LIMIT = {
|
100 |
"gpt-3.5-turbo": 4096,
|
101 |
"gpt-3.5-turbo-0301": 4096,
|
modules/utils.py
CHANGED
@@ -33,7 +33,7 @@ if TYPE_CHECKING:
|
|
33 |
class DataframeData(TypedDict):
|
34 |
headers: List[str]
|
35 |
data: List[List[str | int | bool]]
|
36 |
-
|
37 |
def predict(current_model, *args):
|
38 |
iter = current_model.predict(*args)
|
39 |
for i in iter:
|
@@ -110,6 +110,9 @@ def set_user_identifier(current_model, *args):
|
|
110 |
def set_single_turn(current_model, *args):
|
111 |
current_model.set_single_turn(*args)
|
112 |
|
|
|
|
|
|
|
113 |
|
114 |
def count_token(message):
|
115 |
encoding = tiktoken.get_encoding("cl100k_base")
|
@@ -197,10 +200,13 @@ def convert_asis(userinput):
|
|
197 |
|
198 |
|
199 |
def detect_converted_mark(userinput):
|
200 |
-
|
|
|
|
|
|
|
|
|
|
|
201 |
return True
|
202 |
-
else:
|
203 |
-
return False
|
204 |
|
205 |
|
206 |
def detect_language(code):
|
|
|
33 |
class DataframeData(TypedDict):
|
34 |
headers: List[str]
|
35 |
data: List[List[str | int | bool]]
|
36 |
+
|
37 |
def predict(current_model, *args):
|
38 |
iter = current_model.predict(*args)
|
39 |
for i in iter:
|
|
|
110 |
def set_single_turn(current_model, *args):
|
111 |
current_model.set_single_turn(*args)
|
112 |
|
113 |
+
def handle_file_upload(current_model, *args):
|
114 |
+
return current_model.handle_file_upload(*args)
|
115 |
+
|
116 |
|
117 |
def count_token(message):
|
118 |
encoding = tiktoken.get_encoding("cl100k_base")
|
|
|
200 |
|
201 |
|
202 |
def detect_converted_mark(userinput):
|
203 |
+
try:
|
204 |
+
if userinput.endswith(ALREADY_CONVERTED_MARK):
|
205 |
+
return True
|
206 |
+
else:
|
207 |
+
return False
|
208 |
+
except:
|
209 |
return True
|
|
|
|
|
210 |
|
211 |
|
212 |
def detect_language(code):
|