Module: fireworks_llama_llm

Expand source code

# Copyright (C) 2023-present The Project Contributors
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
#    http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.

from dataclasses import dataclass
import fireworks.client  # noqa
from cl.convince.llms.llama.llama_llm import LlamaLlm
from cl.convince.llms.llm import Llm
from cl.convince.settings.fireworks_settings import FireworksSettings


@dataclass(slots=True, kw_only=True)
class FireworksLlamaLlm(LlamaLlm):
    """Implements LLAMA API running in the Fireworks cloud."""

    model_name: str | None = None
    """Model name in Fireworks format including version if any, defaults to 'llm_id'."""

    max_tokens: int = 4096
    """Maximum number of tokens the model will generate in response to the query."""

    def uncached_completion(self, request_id: str, query: str) -> str:
        """Perform completion without CompletionCache lookup, call completion instead."""

        # Prefix a unique RequestID to the model for audit log purposes and
        # to stop model provider from caching the results
        query_with_request_id = f"RequestID: {request_id}nn{query}"

        model_name = self.model_name if self.model_name is not None else self.llm_id
        prompt = f"""<|begin_of_text|><|start_header_id|>user<|end_header_id|>

{query_with_request_id}<|eot_id|>
<|start_header_id|>assistant<|end_header_id|>"""
        fireworks.client.api_key = FireworksSettings.instance().api_key
        response = fireworks.client.Completion.create(
            model=f"accounts/fireworks/models/{model_name}", prompt=prompt, max_tokens=self.max_tokens
        )
        result = response.choices[0].text
        return result

Classes

class FireworksLlamaLlm (*, llm_id: str = None, model_name: str | None = None, max_tokens: int = 4096)

Implements LLAMA API running in the Fireworks cloud.

Expand source code

@dataclass(slots=True, kw_only=True)
class FireworksLlamaLlm(LlamaLlm):
    """Implements LLAMA API running in the Fireworks cloud."""

    model_name: str | None = None
    """Model name in Fireworks format including version if any, defaults to 'llm_id'."""

    max_tokens: int = 4096
    """Maximum number of tokens the model will generate in response to the query."""

    def uncached_completion(self, request_id: str, query: str) -> str:
        """Perform completion without CompletionCache lookup, call completion instead."""

        # Prefix a unique RequestID to the model for audit log purposes and
        # to stop model provider from caching the results
        query_with_request_id = f"RequestID: {request_id}nn{query}"

        model_name = self.model_name if self.model_name is not None else self.llm_id
        prompt = f"""<|begin_of_text|><|start_header_id|>user<|end_header_id|>

{query_with_request_id}<|eot_id|>
<|start_header_id|>assistant<|end_header_id|>"""
        fireworks.client.api_key = FireworksSettings.instance().api_key
        response = fireworks.client.Completion.create(
            model=f"accounts/fireworks/models/{model_name}", prompt=prompt, max_tokens=self.max_tokens
        )
        result = response.choices[0].text
        return result

Ancestors

LlamaLlm
Llm
LlmKey
KeyMixin
RecordMixin
abc.ABC
typing.Generic

Static methods

def create_prompt_from_messages(messages: list[dict]) -> str: Inherited from: LlamaLlm.create_prompt_from_messages

Transforms list of messages in the following format: [ {“role”: “system”, “content”: “System Prompt”}, {“role”: “user”, “content”: “What is 2 …
def create_prompt_from_messages_v2(messages: list[dict]) -> str: Inherited from: LlamaLlm.create_prompt_from_messages_v2

Transforms list of messages in the following format: [ {“role”: “system”, “content”: “System Prompt”}, {“role”: “user”, “content”: “What is 2 …
def get_key_type() -> Type: Inherited from: LlamaLlm.get_key_type

Return key type even when called from a record.

Fields

var llm_id -> str: Inherited from: LlamaLlm.llm_id

Unique LLM identifier.
var max_tokens -> int: Maximum number of tokens the model will generate in response to the query.
var model_name -> str | None: Model name in Fireworks format including version if any, defaults to ‘llm_id’.

Methods

def completion(self, query: str, *, trial_id: str | int | None = None) -> str: Inherited from: LlamaLlm.completion

Text-in, text-out single query completion without model-specific tags (uses response caching).
def get_key(self) -> LlmKey: Inherited from: LlamaLlm.get_key

Return a new key object whose fields populated from self, do not return self.
def init_all(self) -> None: Inherited from: LlamaLlm.init_all

Invoke ‘init’ for each class in the order from base to derived, then validate against schema.
def uncached_completion(self, request_id: str, query: str) -> str: Inherited from: LlamaLlm.uncached_completion

Perform completion without CompletionCache lookup, call completion instead.