Ishaan Jaff
ffb752d878
fix entrypoint
2024-08-26 20:32:23 -07:00
Krrish Dholakia
b989762bb0
fix(sagemaker.py): support streaming for messages api
...
Fixes https://github.com/BerriAI/litellm/issues/5372
2024-08-26 15:08:08 -07:00
Ishaan Jaff
5c1ebb6ac2
use common folder for cohere
2024-08-26 14:28:50 -07:00
Ishaan Jaff
2481d6007b
refactor cohere to be in a folder
2024-08-26 14:16:25 -07:00
Krrish Dholakia
8df5b25b31
fix(main.py): fix linting errors
2024-08-26 11:44:37 -07:00
Krish Dholakia
bc2a96b2a5
Merge branch 'main' into litellm_vertex_migration
2024-08-24 18:24:19 -07:00
Krrish Dholakia
79bfdb83cc
fix(main.py): cover openai /v1/completions endpoint
2024-08-24 13:25:17 -07:00
Krrish Dholakia
ea2077be22
test(test_router.py): add test to ensure error is correctly re-raised
2024-08-24 10:08:14 -07:00
Krish Dholakia
5eba49c112
Merge pull request #5343 from BerriAI/litellm_sagemaker_chat
...
feat(sagemaker.py): add sagemaker messages api support
2024-08-23 21:00:00 -07:00
Ishaan Jaff
d0c369477b
add mock testing for vertex tts
2024-08-23 18:18:37 -07:00
Ishaan Jaff
39efd58b1e
docs on using vertex tts
2024-08-23 17:57:49 -07:00
Ishaan Jaff
01f766bf89
fix linting
2024-08-23 16:05:31 -07:00
Ishaan Jaff
4e9a44f8d3
fix linting errors
2024-08-23 15:44:31 -07:00
Krrish Dholakia
f7aa787fe6
feat(sagemaker.py): add sagemaker messages api support
...
Closes https://github.com/BerriAI/litellm/issues/2641
Closes https://github.com/BerriAI/litellm/pull/5178
2024-08-23 10:31:35 -07:00
Krish Dholakia
11cbf60e4f
Merge branch 'main' into litellm_azure_batch_apis
2024-08-22 19:07:54 -07:00
Ishaan Jaff
2864d16fa1
Merge branch 'main' into litellm_allow_using_azure_ad_token_auth
2024-08-22 18:21:24 -07:00
Krrish Dholakia
b7f42c96a4
feat(azure.py): support health checking azure deployments
...
Fixes https://github.com/BerriAI/litellm/issues/5279
2024-08-22 16:11:14 -07:00
Ishaan Jaff
b16752f0bc
add new litellm params for client_id, tenant_id etc
2024-08-22 11:37:30 -07:00
Ishaan Jaff
525d152d85
use azure_ad_token_provider to init clients
2024-08-22 11:03:49 -07:00
Krrish Dholakia
900d8ecbf0
feat(factory.py): enable 'user_continue_message' for interweaving user/assistant messages when provider requires it
...
allows bedrock to be used with autogen
2024-08-22 11:03:33 -07:00
Krrish Dholakia
8f306f8e41
fix(cohere_chat.py): support passing 'extra_headers'
...
Fixes https://github.com/BerriAI/litellm/issues/4709
2024-08-22 10:17:36 -07:00
Krrish Dholakia
2dd616bad0
fix(ollama_chat.py): fix passing assistant message with tool call param
...
Fixes https://github.com/BerriAI/litellm/issues/5319
2024-08-22 10:00:03 -07:00
Ishaan Jaff
0435101df4
add multi modal vtx embedding
2024-08-21 15:05:59 -07:00
Ishaan Jaff
be6eb52036
add initial support for multimodal_embedding vertex
2024-08-21 14:29:05 -07:00
Krish Dholakia
85c5d4f8c2
Merge branch 'main' into litellm_fix_azure_api_version
2024-08-20 11:40:53 -07:00
Krrish Dholakia
bc0710a3e6
fix(main.py): response_format typing for acompletion
...
Fixes https://github.com/BerriAI/litellm/issues/5239
2024-08-20 08:14:14 -07:00
Krrish Dholakia
d08479b52c
feat(azure.py): support dynamic api versions
...
Closes https://github.com/BerriAI/litellm/issues/5228
2024-08-19 12:17:43 -07:00
Krish Dholakia
0916197c9d
Merge pull request #5244 from BerriAI/litellm_better_error_logging_sentry
...
refactor: replace .error() with .exception() logging for better debugging on sentry
2024-08-16 19:16:20 -07:00
Krrish Dholakia
9efe9982f5
fix(health_check.py): return 'missing mode' error message, if error with health check, and mode is missing
2024-08-16 17:24:29 -07:00
Krrish Dholakia
2874b94fb1
refactor: replace .error() with .exception() logging for better debugging on sentry
2024-08-16 09:22:47 -07:00
Ishaan Jaff
953a67ba4c
refactor sagemaker to be async
2024-08-15 18:18:02 -07:00
Krrish Dholakia
3026e69926
fix(utils.py): support calling openai models via azure_ai/
2024-08-14 13:41:04 -07:00
Krrish Dholakia
3fd02a1587
fix(main.py): safely fail stream_chunk_builder calls
2024-08-10 10:22:26 -07:00
Krrish Dholakia
5357bd7799
docs(main.py): clarify 'num_retries' usage
2024-08-09 16:57:06 -07:00
Krrish Dholakia
abda7e37ec
fix(utils.py): set max_retries = num_retries, if given
2024-08-09 16:54:54 -07:00
Ishaan Jaff
0561c06bcf
fix cohere / cohere_chat when timeout is None
2024-08-09 12:10:02 -07:00
Krish Dholakia
ab487f0cf9
Merge pull request #5079 from BerriAI/litellm_add_pydantic_model_support
...
feat(utils.py): support passing response_format as pydantic model
2024-08-07 14:43:05 -07:00
Krish Dholakia
78d72acc44
Merge branch 'main' into litellm_anthropic_streaming_tool_call_fix
2024-08-07 14:33:30 -07:00
Krrish Dholakia
661529beb7
fix(main.py): fix linting error for python3.8
2024-08-07 13:21:35 -07:00
Krish Dholakia
77a33baabb
Merge branch 'main' into litellm_add_pydantic_model_support
2024-08-07 13:07:46 -07:00
Krrish Dholakia
ec4051592b
fix(anthropic.py): handle scenario where anthropic returns invalid json string for tool call while streaming
...
Fixes https://github.com/BerriAI/litellm/issues/5063
2024-08-07 09:24:11 -07:00
Ishaan Jaff
7d6d7f2bab
fix use extra headers for open router
2024-08-07 08:15:05 -07:00
Krrish Dholakia
2b132c6bef
feat(utils.py): support passing response_format as pydantic model
...
Related issue - https://github.com/BerriAI/litellm/issues/5074
2024-08-06 18:16:07 -07:00
Krrish Dholakia
8a671aa20b
fix(main.py): log hidden params for text completion calls
2024-08-05 21:26:48 -07:00
Krrish Dholakia
8500f6d087
feat(caching.py): enable caching on provider-specific optional params
...
Closes https://github.com/BerriAI/litellm/issues/5049
2024-08-05 11:18:59 -07:00
Krish Dholakia
88a2428ee6
Merge branch 'main' into litellm_anthropic_api_streaming
2024-08-03 21:16:50 -07:00
Krrish Dholakia
5810708c71
feat(anthropic_adapter.py): support streaming requests for /v1/messages endpoint
...
Fixes https://github.com/BerriAI/litellm/issues/5011
2024-08-03 20:16:19 -07:00
Joe Cheng
fa9298d588
Use correct key name
2024-08-03 11:58:46 -07:00
Joe Cheng
b4df896ac9
Fix tool call coalescing
...
The previous code seemed to assume that the tool call index property
started at 0, but Anthropic sometimes returns them starting at 1.
This was causing an extra null-ish tool call to be materialized.
2024-08-02 13:05:23 -07:00
Joe Cheng
0f301a120f
fix(main.py): Handle bedrock tool calling in stream_chunk_builder
...
Fixes #5022 .
The streaming chunks from Anthropic seem to violate an assumption
that is implicit in the stream_chunk_builder implementation: that
only tool_calls OR function_calls OR content will appear in a
streamed response. The repro in #5022 shows that you can get
content followed by tool calls.
These changes properly handle these combinations by building
separate lists of each type of chunk (note that in theory a chunk
could appear in multiple lists, e.g. both delta.tool_calls and
delta.content being present on one chunk).
2024-08-02 12:41:13 -07:00
