oobabooga
dc92ee774a
Sanitize character name in load_character to prevent path traversal
2026-05-20 12:23:00 -03:00
oobabooga
350eb1780d
Fix thinking channel being lost across tool-call turns ( closes #7578 )
2026-05-20 06:55:50 -07:00
oobabooga
2b07e17c52
Show live generation speed (tokens/s) while generating ( closes #7563 )
2026-05-17 18:37:25 -07:00
oobabooga
d803e29d7f
Fix continue-mode regressions across template families
2026-05-16 13:40:24 -07:00
MeemeeLab
9a396e1749
Fix incorrect prompts generated with continue mode
...
When using the continue function, add_generation_prompt was set to False,
bypassing any template logic that depends on it. This caused missing tokens
or headers that some models inject at the start of an assistant turn.
Gemma 4 was particularly sensitive to this, producing garbled output due
to the missing thought channel header (<|channel>thought\n<channel|>).
Fix by popping the last incomplete assistant message and re-rendering the
prompt with add_generation_prompt=True, then appending the partial content
afterward. This ensures the prompt is structurally identical to normal
generation regardless of the model's template.
GPT-OSS and Seed-OSS thinking block handling is preserved via the existing
fake-message approach, as those models manage thinking content differently.
2026-05-16 13:40:15 -07:00
oobabooga
8e34c7b1b0
Show live context size while generating
2026-05-15 12:45:05 -07:00
oobabooga
2a27ceae93
Hide reasoning and tools controls in chat mode
2026-05-15 11:44:09 -07:00
oobabooga
1e8136dc2a
Small simplifications
2026-05-14 11:21:18 -07:00
oobabooga
58774c8cb1
Also fix streaming leak across all other chat actions
2026-05-14 11:07:59 -07:00
oobabooga
630141912b
Fix streaming output leaking across chats ( closes #7555 )
2026-05-14 08:36:49 -07:00
oobabooga
88d47dc412
Don't clear chat input when no model is loaded ( closes #7542 )
2026-05-07 06:34:10 -07:00
oobabooga
8e0b02a169
Fix Stop button being ignored during tool call approval
2026-05-02 08:26:38 -07:00
oobabooga
54c4feeb4f
Support .jinja/.jinja2 instruction template files in the UI ( closes #7517 )
2026-04-26 18:57:07 -07:00
oobabooga
7338120bff
Fix GPT-OSS channel tokens leaking in UI after tool calls
2026-04-22 18:08:59 -07:00
oobabooga
83347ed7e4
Add preserve_thinking chat template parameter (UI and CLI)
2026-04-22 13:43:56 -07:00
oobabooga
a253d38953
Add tool call confirmation with inline approve/reject buttons
2026-04-22 11:49:39 -07:00
oobabooga
707b5df4ea
Add stdio MCP server support via user_data/mcp.json
2026-04-19 20:47:59 -07:00
oobabooga
78a178de9e
UI: Past chats menu improvement
2026-04-16 03:32:44 -07:00
oobabooga
e4168ee8b8
Fix tool responses with Gemma 4 template ( closes #7498 )
2026-04-15 08:56:57 -07:00
oobabooga
5992fa9044
Fix "Start reply with" crash ( closes #7497 )
2026-04-15 08:12:35 -07:00
oobabooga
f66a01bd06
UI: Add an error message when no model is loaded
2026-04-14 21:05:56 -07:00
oobabooga
a8a0f8dc82
Fix version metadata not syncing on Continue ( closes #7492 )
2026-04-14 23:13:58 -03:00
oobabooga
775c913de2
Fix crash when truncating prompts with tool call messages
2026-04-06 14:13:01 -07:00
oobabooga
b1d06dcf96
UI: Add MCP server support
2026-04-05 19:46:01 -07:00
oobabooga
e0ad4e60df
UI: Fix tool buffer check truncating visible text at end of generation
2026-04-04 09:57:07 -07:00
oobabooga
54b2f39c78
Cleanup modules/chat.py
2026-04-03 22:07:21 -07:00
oobabooga
6e2b70bde6
Add Gemma 4 tool-calling support
2026-04-02 20:26:27 -07:00
oobabooga
bda95172bd
Fix stopping string detection for chromadb/context-1
2026-03-28 06:09:53 -07:00
Phrosty1
bde496ea5d
Fix prompt corruption when continuing with context truncation ( #7439 )
2026-03-22 21:48:56 -03:00
oobabooga
dde1764763
Cleanup modules/chat.py
2026-03-18 21:12:14 -07:00
oobabooga
fef2bd8630
UI: Fix the instruction template delete dialog not appearing
2026-03-17 22:52:32 -07:00
oobabooga
c8bb2129ba
Security: server-side file save roots, image URL SSRF protection, extension allowlist
2026-03-17 22:29:35 -07:00
oobabooga
27a6cdeec1
Fix multi-turn thinking block corruption for Kimi models
2026-03-17 11:31:55 -07:00
oobabooga
c0de1d176c
UI: Add an incognito chat option
2026-03-15 17:57:31 -07:00
oobabooga
5763cab3c4
Fix a crash loading the MiniMax-M2.5 jinja template
2026-03-15 07:13:26 -07:00
oobabooga
573617157a
Optimize tool call detection
...
Avoids templates that don't contain a given necessary keyword
2026-03-14 12:09:41 -07:00
oobabooga
8bff331893
UI: Fix tool call markup flashing before accordion appears during streaming
2026-03-14 09:26:20 -07:00
oobabooga
cb08ba63dc
Fix GPT-OSS channel markup leaking into UI when model skips analysis block
2026-03-14 09:08:05 -07:00
oobabooga
accb2ef661
UI/API: Prevent tool call markup from leaking into streamed UI output ( closes #7427 )
2026-03-14 06:26:47 -07:00
oobabooga
e8d1c66303
Clean up tool calling code
2026-03-13 18:27:01 -07:00
oobabooga
24e7e77b55
Clean up
2026-03-13 12:37:10 -07:00
oobabooga
e0a38da9f3
Improve tool call parsing for Devstral/GPT-OSS and preserve thinking across tool turns
2026-03-13 11:04:06 -03:00
oobabooga
c094bc943c
UI: Skip output extensions on intermediate tool-calling turns
2026-03-12 21:45:38 -07:00
oobabooga
85ec85e569
UI: Fix Continue while in a tool-calling loop, remove the upper limit on number of tool calls
2026-03-12 20:22:35 -07:00
oobabooga
58f26a4cc7
UI: Skip redundant work in chat loop when no tools are selected
2026-03-12 19:18:55 -07:00
oobabooga
286ae475f6
UI: Clean up tool calling code
2026-03-12 22:39:38 -03:00
oobabooga
a09f21b9de
UI: Fix tool calling for GPT-OSS and Continue
2026-03-12 22:17:20 -03:00
oobabooga
3e6bd1a310
UI: Prepend thinking tag when template appends it to prompt
...
Makes Qwen models have a thinking block straight away during streaming.
2026-03-12 14:30:51 -07:00
oobabooga
9a7428b627
UI: Add collapsible accordions for tool calling steps
2026-03-12 14:16:04 -07:00
oobabooga
a916fb0e5c
API: Preserve mid-conversation system message positions
2026-03-12 14:27:24 -03:00
