Litellm dev 12 25 2025 p2 (#7420)

* test: add new test image embedding to base llm unit tests

Addresses https://github.com/BerriAI/litellm/issues/6515

* fix(bedrock/embed/multimodal-embeddings): strip data prefix from image urls for bedrock multimodal embeddings

Fix https://github.com/BerriAI/litellm/issues/6515

* feat: initial commit for fireworks ai audio transcription support

Relevant issue: https://github.com/BerriAI/litellm/issues/7134

* test: initial fireworks ai test

* feat(fireworks_ai/): implemented fireworks ai audio transcription config

* fix(utils.py): register fireworks ai audio transcription config, in config manager

* fix(utils.py): add fireworks ai param translation to 'get_optional_params_transcription'

* refactor(fireworks_ai/): define text completion route with model name handling

moves model name handling to specific fireworks routes, as required by their api

* refactor(fireworks_ai/chat): define transform_Request - allows fixing model if accounts/ is missing

* fix: fix linting errors

* fix: fix linting errors

* fix: fix linting errors

* fix: fix linting errors

* fix(handler.py): fix linting errors

* fix(main.py): fix tgai text completion route

* refactor(together_ai/completion): refactors together ai text completion route to just use provider transform request

* refactor: move test_fine_tuning_api out of local_testing

reduces local testing ci/cd time

litellm/utils.py

@@ -110,10 +110,12 @@ from litellm.router_utils.get_retry_from_policy import (
 from litellm.secret_managers.main import get_secret
 from litellm.types.llms.openai import (
     AllMessageValues,
+    AllPromptValues,
     ChatCompletionAssistantToolCall,
     ChatCompletionNamedToolChoiceParam,
     ChatCompletionToolParam,
     ChatCompletionToolParamFunctionChunk,
+    OpenAITextCompletionUserMessage,
 )
 from litellm.types.rerank import RerankResponse
 from litellm.types.utils import FileTypes  # type: ignore
@@ -169,7 +171,11 @@ from typing import (
 
 from openai import OpenAIError as OriginalError
 
+from litellm.llms.base_llm.audio_transcription.transformation import (
+    BaseAudioTranscriptionConfig,
+)
 from litellm.llms.base_llm.chat.transformation import BaseConfig
+from litellm.llms.base_llm.completion.transformation import BaseTextCompletionConfig
 from litellm.llms.base_llm.embedding.transformation import BaseEmbeddingConfig
 from litellm.llms.base_llm.rerank.transformation import BaseRerankConfig
 
@@ -1825,6 +1831,19 @@ def supports_vision(model: str, custom_llm_provider: Optional[str] = None) -> bo
         return False
 
 
+def supports_embedding_image_input(
+    model: str, custom_llm_provider: Optional[str] = None
+) -> bool:
+    """
+    Check if the given model supports embedding image input and return a boolean value.
+    """
+    return _supports_factory(
+        model=model,
+        custom_llm_provider=custom_llm_provider,
+        key="supports_embedding_image_input",
+    )
+
+
 def supports_parallel_function_calling(model: str):
     """
     Check if the given model supports parallel function calling and return True if it does, False otherwise.
@@ -2081,6 +2100,13 @@ def get_optional_params_transcription(
                     )
             return non_default_params
 
+    provider_config: Optional[BaseAudioTranscriptionConfig] = None
+    if custom_llm_provider is not None:
+        provider_config = ProviderConfigManager.get_provider_audio_transcription_config(
+            model=model,
+            provider=LlmProviders(custom_llm_provider),
+        )
+
     if custom_llm_provider == "openai" or custom_llm_provider == "azure":
         optional_params = non_default_params
     elif custom_llm_provider == "groq":
@@ -2092,6 +2118,15 @@ def get_optional_params_transcription(
             model=model,
             drop_params=drop_params if drop_params is not None else False,
         )
+    elif provider_config is not None:  # handles fireworks ai, and any future providers
+        supported_params = provider_config.get_supported_openai_params(model=model)
+        _check_valid_arg(supported_params=supported_params)
+        optional_params = provider_config.map_openai_params(
+            non_default_params=non_default_params,
+            optional_params=optional_params,
+            model=model,
+            drop_params=drop_params if drop_params is not None else False,
+        )
     for k in passed_params.keys():  # pass additional kwargs without modification
         if k not in default_params.keys():
             optional_params[k] = passed_params[k]
@@ -4432,6 +4467,9 @@ def _get_model_info_helper(  # noqa: PLR0915
                 supports_audio_input=_model_info.get("supports_audio_input", False),
                 supports_audio_output=_model_info.get("supports_audio_output", False),
                 supports_pdf_input=_model_info.get("supports_pdf_input", False),
+                supports_embedding_image_input=_model_info.get(
+                    "supports_embedding_image_input", False
+                ),
                 tpm=_model_info.get("tpm", None),
                 rpm=_model_info.get("rpm", None),
             )
@@ -6000,6 +6038,15 @@ def is_base64_encoded(s: str) -> bool:
         return False
 
 
+def get_base64_str(s: str) -> str:
+    """
+    s: b64str OR data:image/png;base64,b64str
+    """
+    if "," in s:
+        return s.split(",")[1]
+    return s
+
+
 def has_tool_call_blocks(messages: List[AllMessageValues]) -> bool:
     """
     Returns true, if messages has tool call blocks.
@@ -6268,6 +6315,26 @@ class ProviderConfigManager:
             return litellm.InfinityRerankConfig()
         return litellm.CohereRerankConfig()
 
+    @staticmethod
+    def get_provider_audio_transcription_config(
+        model: str,
+        provider: LlmProviders,
+    ) -> Optional[BaseAudioTranscriptionConfig]:
+        if litellm.LlmProviders.FIREWORKS_AI == provider:
+            return litellm.FireworksAIAudioTranscriptionConfig()
+        return None
+
+    @staticmethod
+    def get_provider_text_completion_config(
+        model: str,
+        provider: LlmProviders,
+    ) -> BaseTextCompletionConfig:
+        if LlmProviders.FIREWORKS_AI == provider:
+            return litellm.FireworksAITextCompletionConfig()
+        elif LlmProviders.TOGETHER_AI == provider:
+            return litellm.TogetherAITextCompletionConfig()
+        return litellm.OpenAITextCompletionConfig()
+
 
 def get_end_user_id_for_cost_tracking(
     litellm_params: dict,