db2df94241
Fix litellm flatten_messages_as_text detection ( #659 )
2025-02-15 11:58:39 +01:00
161ef452c3
Merge branch 'main' of github.com:huggingface/smolagents
2025-02-15 11:17:39 +01:00
d34e0c81d9
Fix litellm flatten_messages_as_text detection
2025-02-15 11:17:16 +01:00
a940f42c04
refactor: simplify file type checking from MIME to extension ( #342 )
2025-02-14 18:03:22 +01:00
d33bc2dd9e
Improve missing toolcall exception ( #614 )
2025-02-14 14:49:15 +01:00
d9da0a70ad
fix tool code generation in case of multiline descriptions ( #613 )
2025-02-14 12:28:08 +01:00
cb2218a86f
Fix for MLX Model when stop sequence is followed by any chars ( #623 )
2025-02-14 12:17:56 +01:00
94371331bb
Fix evaluate_condition for non-bool result ( #638 )
2025-02-14 11:21:04 +01:00
1c1418dfa2
Share full agents ( #533 )
...
Co-authored-by: Alex <sysradium@users.noreply.github.com>
Co-authored-by: Parteek <parteekkamboj112@gmail.com>
Co-authored-by: Albert Villanova del Moral <8515462+albertvillanova@users.noreply.github.com>
Co-authored-by: David Berenstein <david.m.berenstein@gmail.com>
2025-02-13 16:24:44 +01:00
cfe599c54a
Test evaluate_condition ( #634 )
2025-02-13 12:03:52 +01:00
41a388dac6
Refactor operations count state setting ( #631 )
2025-02-13 11:22:02 +01:00
1516ce8d74
Move plan user prompt to YAML and test text of plan prompts ( #591 )
2025-02-13 10:13:25 +01:00
5fd0a2e86e
Add support for non-bool comparison operators. ( #612 )
2025-02-12 19:04:31 +01:00
9b96199d00
MLX model support ( #300 )
2025-02-12 18:30:25 +01:00
833aec9198
Refactor and test final answer prompts ( #595 )
2025-02-12 09:30:26 +01:00
a17f915f61
Test agent call with provide_run_summary ( #513 )
2025-02-11 15:51:49 +01:00
9318c8cae3
Move check_module_authorized out of import_module for use in get_safe_module ( #507 )
...
Co-authored-by: Aymeric Roucher <69208727+aymeric-roucher@users.noreply.github.com>
2025-02-11 11:21:43 +01:00
8c6f90cc11
Test evaluate_delete ( #519 )
2025-02-11 11:06:31 +01:00
91badcc054
Fix 429 Client Error: Too Many Requests in CI test ( #599 )
2025-02-11 10:42:27 +01:00
f6b0a12408
Do not pass system prompt to update plan and test plan prompts ( #586 )
2025-02-10 16:23:04 +01:00
75b2a10fbc
Restore missing user prompt for initial facts ( #576 )
...
Co-authored-by: Albert Villanova del Moral <8515462+albertvillanova@users.noreply.github.com>
2025-02-10 10:00:02 +01:00
823b8df0fb
Test E2B Executor ( #540 )
2025-02-07 15:51:14 +01:00
ecabb9ea4f
Delete prompts_path argument and use prompt_templates ( #541 )
2025-02-07 15:29:03 +01:00
127a042cfb
Fix installation with data files ( #536 )
2025-02-07 13:44:42 +01:00
932298696c
Support disabling logging to terminal ( #512 )
2025-02-07 13:44:21 +01:00
c4bd41d39c
Test HfApiModel call with custom_role_conversions ( #517 )
2025-02-07 10:56:42 +01:00
8ba036bba0
Refactor prompts ( #502 )
...
Co-authored-by: Albert Villanova del Moral <8515462+albertvillanova@users.noreply.github.com>
2025-02-06 18:21:46 +01:00
e872d91972
Test PrintContainer ( #515 )
...
* Rename test_python_interpreter to test_local_python_executor
* Test PrintContainer
* Fix CI tests
2025-02-06 14:02:52 +01:00
ce5ede8c12
Suppress terminal logging in CI tests ( #504 )
...
* Test logging to terminal is disabled for testing
* Suppress terminal logging in CI tests
* Pass verbosity_level to test terminal logging output
* Refactor
2025-02-05 18:48:39 +01:00
66b76200e7
Initialize step_number attribute ( #497 )
...
* Test step_number attribute is initialized
* Initialize step_number attribute
* Fix style
2025-02-05 11:55:37 +01:00
26c733c745
Fix undefined `step_number` in single_step run ( #495 )
...
* Fix undefined step_number
* Test step_number is defined after single_step run
---------
Co-authored-by: Albert Villanova del Moral <8515462+albertvillanova@users.noreply.github.com>
2025-02-05 09:00:19 +01:00
8b02821ac2
Open Deep Research ( #317 )
...
* Create an open Deep Research
Co-authored-by: Albert Villanova del Moral <8515462+albertvillanova@users.noreply.github.com>
---------
Co-authored-by: Albert Villanova del Moral <8515462+albertvillanova@users.noreply.github.com>
2025-02-04 17:13:21 +01:00
d69ae028fa
Simplify managed agents ( #484 )
...
* Simplify managed agents
2025-02-03 21:19:54 +01:00
183869de04
Fix SpeechToTextTool new instance ( #478 )
...
* Test SpeechToTextTool new instance
* Fix SpeechToTextTool new method
2025-02-03 11:42:03 +01:00
cedf63cde7
Test get_clean_message_list ( #448 )
...
* Test get_clean_message_list
* Test get_clean_message_list
2025-01-31 13:47:02 +01:00
f1a9b83443
Improve code execution logs in case of error by showing print outputs ( #446 )
...
* Improve code execution logs in case of error by still showing print outputs
* Improve action step testing
* Number steps starting at 1
2025-01-31 12:34:32 +01:00
10407813e8
Fix get safe module ( #405 )
...
* expose get_safe_module lazy loading issue
* fix get_safe_module lazy imports
* add six as explicit tests dependency
* log dangerous attributes
* refactor get_safe_module test
2025-01-31 08:25:22 +01:00
181a500c5d
Fix and test MemoryStep ( #432 )
...
* Test MemoryStep
* Remove unused MemoryStep.raw attribute
2025-01-30 21:24:39 +01:00
42d97716fe
Fix MultiStepAgent.planning_step message content ( #437 )
...
* Fix MultiStepAgent.planning_step message content
2025-01-30 21:18:30 +01:00
fa1f8d0154
Add CLI for smolagents ( #431 )
...
---------
Co-authored-by: Merve Noyan <merveenoyan@gmail.com>
2025-01-30 15:56:50 +01:00
595781b82d
Test memory ( #428 )
...
* Test memory
* Add memory tests to CI
2025-01-30 13:51:18 +01:00
20383f0914
Fix regressions, refactor logging and improve replay function ( #419 )
...
* Refactor logging and improve replay function
* Fix test
* Remode unnecessary dict method
* Remove logger import
---------
Co-authored-by: Albert Villanova del Moral <8515462+albertvillanova@users.noreply.github.com>
2025-01-30 11:32:28 +01:00
d3912c70cf
Support third-party Inference providers in `HfApiModel` ( #422 )
...
* Add `provider` param to `HfApiModel `, update guided_tour.md
---------
Co-authored-by: Aymeric <aymeric.roucher@gmail.com>
2025-01-30 01:03:09 +01:00
c6eb9526f1
Untangling Logging ( #316 )
...
* Deep refacto that disentangled logging logic between logging (informing the user via console outputs) and memory (storing what happened in the agent).
---------
Co-authored-by: benediktstroebl <50178209+benediktstroebl@users.noreply.github.com>
Co-authored-by: Aymeric <aymeric.roucher@gmail.com>
2025-01-29 16:27:04 +01:00
7a2f300b2a
Add test for DDGS Tool kwargs ( #387 )
2025-01-28 11:02:21 +01:00
ad5f84b101
Fix blocking of os in authorized imports ( #386 )
2025-01-28 11:00:43 +01:00
a4f89b68b2
Add tool saving test ( #389 )
...
* Add tool saving test
* Format
2025-01-28 10:58:53 +01:00
cc21135188
Test import without extras ( #370 )
...
* Test import without extras
* Add import test to CI
* Use uv run
2025-01-28 10:57:14 +01:00
49c34f625c
Gradio chatbot: step duration, number, token count, support nested thoughts ( #384 )
...
* Add enhanced Gradio UI with nested agents calls, execution-logs, errors
- Add virtual separation between steps
- Highlight final answer as required in internal discussion
- Show step numbers and token counts
- Include step duration tracking
- Improve message display structure
---------
Co-authored-by: Yuvraj Sharma <48665385+yvrjsharma@users.noreply.github.com>
2025-01-28 09:24:13 +01:00
4579a6f7cc
TransformersModel auto-detects VLMs ( #378 )
...
* TransformersModel auto-detects VLMs
2025-01-27 20:09:14 +01:00
Aymeric Roucher
Aymeric
kingdomad
Graham
Alex
Joe
Albert Villanova del Moral
Billy Neuson
CalOmnie
NeuroWhAI
Taha Yassine
Antoine Jeannot
Julien Chaumond
Clémentine Fourrier