Skip to content

Ollama

OllamaMultiModal #

Bases: MultiModalLLM

Source code in llama-index-integrations/multi_modal_llms/llama-index-multi-modal-llms-ollama/llama_index/multi_modal_llms/ollama/base.py
 52
 53
 54
 55
 56
 57
 58
 59
 60
 61
 62
 63
 64
 65
 66
 67
 68
 69
 70
 71
 72
 73
 74
 75
 76
 77
 78
 79
 80
 81
 82
 83
 84
 85
 86
 87
 88
 89
 90
 91
 92
 93
 94
 95
 96
 97
 98
 99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
class OllamaMultiModal(MultiModalLLM):
    base_url: str = Field(
        default="http://localhost:11434",
        description="Base url the model is hosted under.",
    )
    model: str = Field(description="The MultiModal Ollama model to use.")
    temperature: float = Field(
        default=0.75,
        description="The temperature to use for sampling.",
        ge=0.0,
        le=1.0,
    )
    context_window: int = Field(
        default=DEFAULT_CONTEXT_WINDOW,
        description="The maximum number of context tokens for the model.",
        gt=0,
    )
    request_timeout: Optional[float] = Field(
        default=60.0,
        description="The timeout for making http request to Ollama API server",
    )
    additional_kwargs: Dict[str, Any] = Field(
        default_factory=dict,
        description="Additional model parameters for the Ollama API.",
    )
    _client: Client = PrivateAttr()
    _aclient: AsyncClient = PrivateAttr()

    def __init__(self, **kwargs: Any) -> None:
        """Init params and ollama client."""
        super().__init__(**kwargs)
        self._client = Client(host=self.base_url, timeout=self.request_timeout)
        self._aclient = AsyncClient(host=self.base_url, timeout=self.request_timeout)

    @classmethod
    def class_name(cls) -> str:
        return "Ollama_multi_modal_llm"

    @property
    def metadata(self) -> MultiModalLLMMetadata:
        """LLM metadata."""
        return MultiModalLLMMetadata(
            context_window=self.context_window,
            num_output=DEFAULT_NUM_OUTPUTS,
            model_name=self.model,
            is_chat_model=True,  # Ollama supports chat API for all models
        )

    @property
    def _model_kwargs(self) -> Dict[str, Any]:
        base_kwargs = {
            "temperature": self.temperature,
            "num_ctx": self.context_window,
        }
        return {
            **base_kwargs,
            **self.additional_kwargs,
        }

    def chat(self, messages: Sequence[ChatMessage], **kwargs: Any) -> ChatResponse:
        """Chat."""
        ollama_messages = _messages_to_dicts(messages)
        response = self._client.chat(
            model=self.model, messages=ollama_messages, stream=False, **kwargs
        )
        return ChatResponse(
            message=ChatMessage(
                content=response["message"]["content"],
                role=MessageRole(response["message"]["role"]),
                additional_kwargs=get_additional_kwargs(response, ("message",)),
            ),
            raw=response["message"],
            additional_kwargs=get_additional_kwargs(response, ("message",)),
        )

    def stream_chat(
        self,
        messages: Sequence[ChatMessage],
        **kwargs: Any,
    ) -> ChatResponseGen:
        """Stream chat."""
        ollama_messages = _messages_to_dicts(messages)
        response = self._client.chat(
            model=self.model, messages=ollama_messages, stream=True, **kwargs
        )
        text = ""
        for chunk in response:
            if "done" in chunk and chunk["done"]:
                break
            message = chunk["message"]
            delta = message.get("content")
            text += delta
            yield ChatResponse(
                message=ChatMessage(
                    content=text,
                    role=MessageRole(message["role"]),
                    additional_kwargs=get_additional_kwargs(
                        message, ("content", "role")
                    ),
                ),
                delta=delta,
                raw=message,
                additional_kwargs=get_additional_kwargs(chunk, ("message",)),
            )

    def complete(
        self,
        prompt: str,
        image_documents: Sequence[ImageNode],
        formatted: bool = False,
        **kwargs: Any,
    ) -> CompletionResponse:
        """Complete."""
        response = self._client.generate(
            model=self.model,
            prompt=prompt,
            images=image_documents_to_base64(image_documents),
            stream=False,
            options=self._model_kwargs,
            **kwargs,
        )
        return CompletionResponse(
            text=response["response"],
            raw=response,
            additional_kwargs=get_additional_kwargs(response, ("response",)),
        )

    def stream_complete(
        self,
        prompt: str,
        image_documents: Sequence[ImageNode],
        formatted: bool = False,
        **kwargs: Any,
    ) -> CompletionResponseGen:
        """Stream complete."""
        response = self._client.generate(
            model=self.model,
            prompt=prompt,
            images=image_documents_to_base64(image_documents),
            stream=True,
            options=self._model_kwargs,
            **kwargs,
        )
        text = ""
        for chunk in response:
            if "done" in chunk and chunk["done"]:
                break
            delta = chunk.get("response")
            text += delta
            yield CompletionResponse(
                text=str(chunk["response"]),
                delta=delta,
                raw=chunk,
                additional_kwargs=get_additional_kwargs(chunk, ("response",)),
            )

    async def acomplete(
        self,
        prompt: str,
        image_documents: Sequence[ImageNode],
        **kwargs: Any,
    ) -> CompletionResponse:
        """Async complete."""
        response = await self._aclient.generate(
            model=self.model,
            prompt=prompt,
            images=image_documents_to_base64(image_documents),
            stream=False,
            options=self._model_kwargs,
            **kwargs,
        )
        return CompletionResponse(
            text=response["response"],
            raw=response,
            additional_kwargs=get_additional_kwargs(response, ("response",)),
        )

    async def achat(
        self,
        messages: Sequence[ChatMessage],
        **kwargs: Any,
    ) -> ChatResponse:
        """Async chat."""
        ollama_messages = _messages_to_dicts(messages)
        response = await self._aclient.chat(
            model=self.model, messages=ollama_messages, stream=False, **kwargs
        )
        return ChatResponse(
            message=ChatMessage(
                content=response["message"]["content"],
                role=MessageRole(response["message"]["role"]),
                additional_kwargs=get_additional_kwargs(response, ("message",)),
            ),
            raw=response["message"],
            additional_kwargs=get_additional_kwargs(response, ("message",)),
        )

    async def astream_complete(
        self,
        prompt: str,
        image_documents: Sequence[ImageNode],
        **kwargs: Any,
    ) -> CompletionResponseAsyncGen:
        """Async stream complete."""
        async for chunk in self._aclient.generate(
            model=self.model,
            prompt=prompt,
            images=image_documents_to_base64(image_documents),
            stream=True,
            options=self._model_kwargs,
            **kwargs,
        ):
            if "done" in chunk and chunk["done"]:
                break
            delta = chunk.get("response")
            yield CompletionResponse(
                text=str(chunk["response"]),
                delta=delta,
                raw=chunk,
                additional_kwargs=get_additional_kwargs(chunk, ("response",)),
            )

    async def astream_chat(
        self,
        messages: Sequence[ChatMessage],
        **kwargs: Any,
    ) -> ChatResponseAsyncGen:
        """Async stream chat."""
        ollama_messages = _messages_to_dicts(messages)
        async for chunk in self._aclient.chat(
            model=self.model, messages=ollama_messages, stream=True, **kwargs
        ):
            if "done" in chunk and chunk["done"]:
                break
            message = chunk["message"]
            delta = message.get("content")
            yield ChatResponse(
                message=ChatMessage(
                    content=message["content"],
                    role=MessageRole(message["role"]),
                    additional_kwargs=get_additional_kwargs(
                        message, ("content", "role")
                    ),
                ),
                delta=delta,
                raw=message,
                additional_kwargs=get_additional_kwargs(chunk, ("message",)),
            )

metadata property #

metadata: MultiModalLLMMetadata

LLM metadata.

chat #

chat(messages: Sequence[ChatMessage], **kwargs: Any) -> ChatResponse

Chat.

Source code in llama-index-integrations/multi_modal_llms/llama-index-multi-modal-llms-ollama/llama_index/multi_modal_llms/ollama/base.py
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
def chat(self, messages: Sequence[ChatMessage], **kwargs: Any) -> ChatResponse:
    """Chat."""
    ollama_messages = _messages_to_dicts(messages)
    response = self._client.chat(
        model=self.model, messages=ollama_messages, stream=False, **kwargs
    )
    return ChatResponse(
        message=ChatMessage(
            content=response["message"]["content"],
            role=MessageRole(response["message"]["role"]),
            additional_kwargs=get_additional_kwargs(response, ("message",)),
        ),
        raw=response["message"],
        additional_kwargs=get_additional_kwargs(response, ("message",)),
    )

stream_chat #

stream_chat(messages: Sequence[ChatMessage], **kwargs: Any) -> ChatResponseGen

Stream chat.

Source code in llama-index-integrations/multi_modal_llms/llama-index-multi-modal-llms-ollama/llama_index/multi_modal_llms/ollama/base.py
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
def stream_chat(
    self,
    messages: Sequence[ChatMessage],
    **kwargs: Any,
) -> ChatResponseGen:
    """Stream chat."""
    ollama_messages = _messages_to_dicts(messages)
    response = self._client.chat(
        model=self.model, messages=ollama_messages, stream=True, **kwargs
    )
    text = ""
    for chunk in response:
        if "done" in chunk and chunk["done"]:
            break
        message = chunk["message"]
        delta = message.get("content")
        text += delta
        yield ChatResponse(
            message=ChatMessage(
                content=text,
                role=MessageRole(message["role"]),
                additional_kwargs=get_additional_kwargs(
                    message, ("content", "role")
                ),
            ),
            delta=delta,
            raw=message,
            additional_kwargs=get_additional_kwargs(chunk, ("message",)),
        )

complete #

complete(prompt: str, image_documents: Sequence[ImageNode], formatted: bool = False, **kwargs: Any) -> CompletionResponse

Complete.

Source code in llama-index-integrations/multi_modal_llms/llama-index-multi-modal-llms-ollama/llama_index/multi_modal_llms/ollama/base.py
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
def complete(
    self,
    prompt: str,
    image_documents: Sequence[ImageNode],
    formatted: bool = False,
    **kwargs: Any,
) -> CompletionResponse:
    """Complete."""
    response = self._client.generate(
        model=self.model,
        prompt=prompt,
        images=image_documents_to_base64(image_documents),
        stream=False,
        options=self._model_kwargs,
        **kwargs,
    )
    return CompletionResponse(
        text=response["response"],
        raw=response,
        additional_kwargs=get_additional_kwargs(response, ("response",)),
    )

stream_complete #

stream_complete(prompt: str, image_documents: Sequence[ImageNode], formatted: bool = False, **kwargs: Any) -> CompletionResponseGen

Stream complete.

Source code in llama-index-integrations/multi_modal_llms/llama-index-multi-modal-llms-ollama/llama_index/multi_modal_llms/ollama/base.py
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
def stream_complete(
    self,
    prompt: str,
    image_documents: Sequence[ImageNode],
    formatted: bool = False,
    **kwargs: Any,
) -> CompletionResponseGen:
    """Stream complete."""
    response = self._client.generate(
        model=self.model,
        prompt=prompt,
        images=image_documents_to_base64(image_documents),
        stream=True,
        options=self._model_kwargs,
        **kwargs,
    )
    text = ""
    for chunk in response:
        if "done" in chunk and chunk["done"]:
            break
        delta = chunk.get("response")
        text += delta
        yield CompletionResponse(
            text=str(chunk["response"]),
            delta=delta,
            raw=chunk,
            additional_kwargs=get_additional_kwargs(chunk, ("response",)),
        )

acomplete async #

acomplete(prompt: str, image_documents: Sequence[ImageNode], **kwargs: Any) -> CompletionResponse

Async complete.

Source code in llama-index-integrations/multi_modal_llms/llama-index-multi-modal-llms-ollama/llama_index/multi_modal_llms/ollama/base.py
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
async def acomplete(
    self,
    prompt: str,
    image_documents: Sequence[ImageNode],
    **kwargs: Any,
) -> CompletionResponse:
    """Async complete."""
    response = await self._aclient.generate(
        model=self.model,
        prompt=prompt,
        images=image_documents_to_base64(image_documents),
        stream=False,
        options=self._model_kwargs,
        **kwargs,
    )
    return CompletionResponse(
        text=response["response"],
        raw=response,
        additional_kwargs=get_additional_kwargs(response, ("response",)),
    )

achat async #

achat(messages: Sequence[ChatMessage], **kwargs: Any) -> ChatResponse

Async chat.

Source code in llama-index-integrations/multi_modal_llms/llama-index-multi-modal-llms-ollama/llama_index/multi_modal_llms/ollama/base.py
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
async def achat(
    self,
    messages: Sequence[ChatMessage],
    **kwargs: Any,
) -> ChatResponse:
    """Async chat."""
    ollama_messages = _messages_to_dicts(messages)
    response = await self._aclient.chat(
        model=self.model, messages=ollama_messages, stream=False, **kwargs
    )
    return ChatResponse(
        message=ChatMessage(
            content=response["message"]["content"],
            role=MessageRole(response["message"]["role"]),
            additional_kwargs=get_additional_kwargs(response, ("message",)),
        ),
        raw=response["message"],
        additional_kwargs=get_additional_kwargs(response, ("message",)),
    )

astream_complete async #

astream_complete(prompt: str, image_documents: Sequence[ImageNode], **kwargs: Any) -> CompletionResponseAsyncGen

Async stream complete.

Source code in llama-index-integrations/multi_modal_llms/llama-index-multi-modal-llms-ollama/llama_index/multi_modal_llms/ollama/base.py
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
async def astream_complete(
    self,
    prompt: str,
    image_documents: Sequence[ImageNode],
    **kwargs: Any,
) -> CompletionResponseAsyncGen:
    """Async stream complete."""
    async for chunk in self._aclient.generate(
        model=self.model,
        prompt=prompt,
        images=image_documents_to_base64(image_documents),
        stream=True,
        options=self._model_kwargs,
        **kwargs,
    ):
        if "done" in chunk and chunk["done"]:
            break
        delta = chunk.get("response")
        yield CompletionResponse(
            text=str(chunk["response"]),
            delta=delta,
            raw=chunk,
            additional_kwargs=get_additional_kwargs(chunk, ("response",)),
        )

astream_chat async #

astream_chat(messages: Sequence[ChatMessage], **kwargs: Any) -> ChatResponseAsyncGen

Async stream chat.

Source code in llama-index-integrations/multi_modal_llms/llama-index-multi-modal-llms-ollama/llama_index/multi_modal_llms/ollama/base.py
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
async def astream_chat(
    self,
    messages: Sequence[ChatMessage],
    **kwargs: Any,
) -> ChatResponseAsyncGen:
    """Async stream chat."""
    ollama_messages = _messages_to_dicts(messages)
    async for chunk in self._aclient.chat(
        model=self.model, messages=ollama_messages, stream=True, **kwargs
    ):
        if "done" in chunk and chunk["done"]:
            break
        message = chunk["message"]
        delta = message.get("content")
        yield ChatResponse(
            message=ChatMessage(
                content=message["content"],
                role=MessageRole(message["role"]),
                additional_kwargs=get_additional_kwargs(
                    message, ("content", "role")
                ),
            ),
            delta=delta,
            raw=message,
            additional_kwargs=get_additional_kwargs(chunk, ("message",)),
        )