not-again-ai 0.8.2__tar.gz → 0.9.0__tar.gz

{not_again_ai-0.8.2 → not_again_ai-0.9.0}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: not-again-ai
-Version: 0.8.2
+Version: 0.9.0
 Summary: Designed to once and for all collect all the little things that come up over and over again in AI projects and put them in one place.
 Home-page: https://github.com/DaveCoDev/not-again-ai
 License: MIT
@@ -18,6 +18,7 @@ Classifier: Programming Language :: Python :: 3.11
 Classifier: Programming Language :: Python :: 3.12
 Classifier: Typing :: Typed
 Provides-Extra: llm
+Provides-Extra: local-llm
 Provides-Extra: statistics
 Provides-Extra: viz
 Requires-Dist: numpy (==1.26.4) ; extra == "statistics" or extra == "viz"
@@ -29,6 +30,7 @@ Requires-Dist: scikit-learn (==1.4.2) ; extra == "statistics"
 Requires-Dist: scipy (==1.13.0) ; extra == "statistics"
 Requires-Dist: seaborn (==0.13.2) ; extra == "viz"
 Requires-Dist: tiktoken (==0.7.0) ; extra == "llm"
+Requires-Dist: transformers (==4.41.1) ; extra == "local-llm"
 Project-URL: Documentation, https://github.com/DaveCoDev/not-again-ai
 Project-URL: Repository, https://github.com/DaveCoDev/not-again-ai
 Description-Content-Type: text/markdown
@@ -59,9 +61,11 @@ Requires: Python 3.11, or 3.12
 Install the entire package from [PyPI](https://pypi.org/project/not-again-ai/) with: 
 
 ```bash
-$ pip install not_again_ai[llm,statistics,viz]
+$ pip install not_again_ai[llm,local_llm,statistics,viz]
 ```
 
+Note that local LLM requires separate installations and will not work out of the box due to how hardware dependent it is. Be sure to check the [notebooks](notebooks/local_llm/) for more details.
+
 The package is split into subpackages, so you can install only the parts you need. See the **[notebooks](notebooks)** for examples.
 * **Base only**: `pip install not_again_ai`
 * **LLM**: `pip install not_again_ai[llm]`
@@ -77,7 +81,9 @@ The package is split into subpackages, so you can install only the parts you nee
             ...
             Environment="OLLAMA_HOST=0.0.0.0"
             ```
-        Now ollama will be available at `http://<local_address>:11434`
+        1. Ollama will be available at `http://<local_address>:11434`
+* **Local LLM**: `pip install not_again_ai[llm_local]`
+    - Most of this package is hardware dependent so this only installs some generic dependencies. Be sure to check the [notebooks](notebooks/local_llm/) for more details on what is available and how to install it.
 * **Statistics**: `pip install not_again_ai[statistics]`
 * **Visualization**: `pip install not_again_ai[viz]`
 

{not_again_ai-0.8.2 → not_again_ai-0.9.0}/README.md

@@ -24,9 +24,11 @@ Requires: Python 3.11, or 3.12
 Install the entire package from [PyPI](https://pypi.org/project/not-again-ai/) with: 
 
 ```bash
-$ pip install not_again_ai[llm,statistics,viz]
+$ pip install not_again_ai[llm,local_llm,statistics,viz]
 ```
 
+Note that local LLM requires separate installations and will not work out of the box due to how hardware dependent it is. Be sure to check the [notebooks](notebooks/local_llm/) for more details.
+
 The package is split into subpackages, so you can install only the parts you need. See the **[notebooks](notebooks)** for examples.
 * **Base only**: `pip install not_again_ai`
 * **LLM**: `pip install not_again_ai[llm]`
@@ -42,7 +44,9 @@ The package is split into subpackages, so you can install only the parts you nee
             ...
             Environment="OLLAMA_HOST=0.0.0.0"
             ```
-        Now ollama will be available at `http://<local_address>:11434`
+        1. Ollama will be available at `http://<local_address>:11434`
+* **Local LLM**: `pip install not_again_ai[llm_local]`
+    - Most of this package is hardware dependent so this only installs some generic dependencies. Be sure to check the [notebooks](notebooks/local_llm/) for more details on what is available and how to install it.
 * **Statistics**: `pip install not_again_ai[statistics]`
 * **Visualization**: `pip install not_again_ai[viz]`
 

{not_again_ai-0.8.2 → not_again_ai-0.9.0}/pyproject.toml

@@ -1,6 +1,6 @@
 [tool.poetry]
 name = "not-again-ai"
-version = "0.8.2"
+version = "0.9.0"
 description = "Designed to once and for all collect all the little things that come up over and over again in AI projects and put them in one place."
 authors = ["DaveCoDev <dave.co.dev@gmail.com>"]
 license = "MIT"
@@ -36,14 +36,17 @@ scipy = { version = "==1.13.0", optional = true }
 scikit-learn = { version = "==1.4.2", optional = true }
 seaborn = { version = "==0.13.2", optional = true }
 tiktoken = { version = "==0.7.0", optional = true }
+transformers = { version = "==4.41.1", optional = true }
 
 [tool.poetry.extras]
 llm = ["ollama", "openai", "python-liquid", "tiktoken"]
+local_llm = ["transformers"]
 statistics = ["numpy", "scikit-learn", "scipy"]
 viz = ["numpy", "pandas", "seaborn"]
 
 [tool.poetry.dev-dependencies]
 ipykernel = "*"
+ipywidgets = "*"
 
 [tool.poetry.group.nox.dependencies]
 nox-poetry = "*"

not_again_ai-0.9.0/src/not_again_ai/local_llm/huggingface/chat_completion.py

@@ -0,0 +1,59 @@
+from pathlib import Path
+from typing import Any
+
+from PIL import Image
+
+
+def chat_completion_image(
+    messages: list[dict[str, str]],
+    images: list[Path] | None,
+    model_processor: tuple[Any, Any],
+    max_tokens: int | None = None,
+    temperature: float = 0.7,
+) -> dict[str, Any]:
+    """A wrapper around ision language model inference for multimodal language models from huggingface.
+
+    Args:
+        messages (list[dict[str, str]]): A list of messages to send to the model.
+        images (list[Path] | None): A list of image paths to send to the model.
+        model_processor (tuple[Any, Any]): A tuple containing the model and processor objects.
+        max_tokens (int, optional): The maximum number of tokens to generate. Defaults to None.
+        temperature (float, optional): The temperature of the model. Increasing the temperature will make the model answer more creatively. Defaults to 0.7.
+
+    Returns:
+        dict[str, Any]: A dictionary with the following keys
+            message (str): The content of the generated assistant message.
+            completion_tokens (int): The number of tokens used by the model to generate the completion.
+    """
+
+    model, processor = model_processor
+
+    prompt = processor.tokenizer.apply_chat_template(messages, tokenize=False, add_generation_prompt=True)
+
+    if images:
+        image_objects = [Image.open(image) for image in images]
+        inputs = processor(prompt, image_objects, return_tensors="pt").to("cuda:0")
+    else:
+        inputs = processor(prompt, return_tensors="pt").to("cuda:0")
+
+    generation_args = {
+        "max_new_tokens": max_tokens,
+        "temperature": temperature,
+        "num_beams": 1,
+        "do_sample": True,
+    }
+
+    generate_ids = model.generate(**inputs, eos_token_id=processor.tokenizer.eos_token_id, **generation_args)
+
+    # Remove input tokens
+    generate_ids = generate_ids[:, inputs["input_ids"].shape[1] :]
+
+    # Get the number of generated tokens
+    completion_tokens = generate_ids.shape[1]
+
+    response = processor.batch_decode(generate_ids, skip_special_tokens=True, clean_up_tokenization_spaces=False)
+
+    response_data: dict[str, Any] = {}
+    response_data["message"] = response[0]
+    response_data["completion_tokens"] = completion_tokens
+    return response_data

not_again_ai-0.9.0/src/not_again_ai/local_llm/huggingface/helpers.py

@@ -0,0 +1,23 @@
+from typing import Any
+
+from transformers import AutoModelForCausalLM, AutoProcessor
+
+
+def load_model(model_id: str, device_map: str = "cuda", trust_remote_code: bool = True) -> Any:
+    """Load a model from Hugging Face."""
+    model = AutoModelForCausalLM.from_pretrained(
+        model_id,
+        device_map=device_map,
+        trust_remote_code=trust_remote_code,
+        torch_dtype="auto",
+    )
+    return model
+
+
+def load_processor(model_id: str, trust_remote_code: bool = True) -> Any:
+    """Load a processor from Hugging Face. This is typically used for multimodal language models."""
+    processor = AutoProcessor.from_pretrained(
+        model_id,
+        trust_remote_code=trust_remote_code,
+    )
+    return processor