ys/GenericAgent
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

restore context_win=10000 max_tokens=6144

Browse Source
This commit is contained in:
Liang Jiaqing
2026-03-06 21:46:41 +08:00
parent 959cf6a5a6
commit 2156bc6fc6
1 changed files with 2 additions and 2 deletions
Download Patch File Download Diff File Expand all files Collapse all files

4
sidercall.py
View File

@@ -37,7 +37,7 @@ class SiderLLMSession:
return full_text return full_text
class ClaudeSession: class ClaudeSession:
def __init__(self, api_key, api_base, model="claude-opus", context_win=9000): def __init__(self, api_key, api_base, model="claude-opus", context_win=10000):
self.api_key, self.api_base, self.default_model, self.context_win = api_key, api_base.rstrip('/'), model, context_win self.api_key, self.api_base, self.default_model, self.context_win = api_key, api_base.rstrip('/'), model, context_win
self.raw_msgs, self.lock = [], threading.Lock() self.raw_msgs, self.lock = [], threading.Lock()
def _trim_messages(self, messages): def _trim_messages(self, messages):
@@ -51,7 +51,7 @@ class ClaudeSession:
else: break else: break
if current > self.context_win * 3.6: print(f'[DEBUG] {len(result)} contexts, whole length {current//4} tokens.') if current > self.context_win * 3.6: print(f'[DEBUG] {len(result)} contexts, whole length {current//4} tokens.')
return result[::-1] or messages[-2:] return result[::-1] or messages[-2:]
def raw_ask(self, messages, model=None, temperature=0.5, max_tokens=4096): def raw_ask(self, messages, model=None, temperature=0.5, max_tokens=6144):
model = model or self.default_model model = model or self.default_model
headers = {"x-api-key": self.api_key, "Content-Type": "application/json", "anthropic-version": "2023-06-01"} headers = {"x-api-key": self.api_key, "Content-Type": "application/json", "anthropic-version": "2023-06-01"}
payload = {"model": model, "messages": messages, "temperature": temperature, "max_tokens": max_tokens, "stream": True} payload = {"model": model, "messages": messages, "temperature": temperature, "max_tokens": max_tokens, "stream": True}