LOGISTAR — Scheda Oggetto

Oggetto salvato correttamente.

Dati oggetto

URL

Apri

Titolo / Descrizione

Sito / Azienda

Estratto / Body

&lt;p&gt;&lt;a href=&quot;https://www.kdnuggets.com/2023/05/ultimate-opensource-large-language-model-ecosystem.html&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot;&gt;GPT4ALL&lt;/a&gt; is a free and open-source AI Playground that can be run locally on Windows, Mac, and Linux computers without requiring an internet connection or a GPU. It offers users access to various state-of-the-art language models through a simple two-step process. All you need to do is to download and install the application and then download your preferred model.&lt;/p&gt;
&lt;p&gt;&lt;a href=&quot;https://github.com/nomic-ai/gpt4all&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot;&gt;GPT4ALL&lt;/a&gt; is a project that provides everything you need to work with state-of-the-art natural language models. You can access open source models and datasets, train and run them with the provided code, use a web interface or a desktop app to interact with them, connect to the Langchain Backend for distributed computing, and use the Python API for easy integration.&lt;/p&gt;
&lt;p&gt;Apache-2 licensed GPT4All-J chatbot was recently launched by the developers, which was trained on a vast, curated corpus of assistant interactions, comprising word problems, multi-turn dialogues, code, poems, songs, and stories. To make it more accessible, they have also released Python bindings and a Chat UI, enabling virtually anyone to run the model on CPU.&amp;nbsp;&lt;/p&gt;
&lt;p&gt;You can try it yourself by installing native chat-client on your desktop.&amp;nbsp;&lt;/p&gt;
&lt;ul&gt;
&lt;li&gt;&lt;a href=&quot;https://gpt4all.io/installers/gpt4all-installer-darwin.dmg&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot;&gt;Mac/OSX&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;&lt;a href=&quot;https://gpt4all.io/installers/gpt4all-installer-win64.exe&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot;&gt;Windows&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;&lt;a href=&quot;https://gpt4all.io/installers/gpt4all-installer-linux.run&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot;&gt;Ubuntu&lt;/a&gt;&lt;/li&gt;
&lt;/ul&gt;
&lt;p&gt;After that, run the GPT4ALL program and download the model of your choice. You can also download models manually &lt;a href=&quot;https://github.com/nomic-ai/gpt4all-chat#manual-download-of-models&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot;&gt;here&lt;/a&gt; and install them in the location indicated by the model download dialog in the GUI.&lt;/p&gt;
&lt;h1&gt;GPT4ALL Python Client&lt;/h1&gt;
&lt;p&gt;&amp;nbsp;&lt;/p&gt;
&lt;p&gt;&amp;nbsp;&lt;/p&gt;
&lt;p&gt;&amp;nbsp;&lt;/p&gt;
&lt;p&gt;The GPT4ALL comes with &lt;a href=&quot;https://github.com/nomic-ai/pyllamacpp&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot;&gt;Python&lt;/a&gt;, &lt;a href=&quot;https://github.com/nomic-ai/gpt4all-ts&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot;&gt;TypeScript&lt;/a&gt;, &lt;a href=&quot;https://github.com/nomic-ai/gpt4all-ui&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot;&gt;Web Chat interface&lt;/a&gt;, and &lt;a href=&quot;https://python.langchain.com/en/latest/modules/models/llms/integrations/gpt4all.html&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot;&gt;Langchain&lt;/a&gt; backend.&amp;nbsp;&amp;nbsp;&lt;/p&gt;
&lt;p&gt;&amp;nbsp;&lt;/p&gt;
&lt;p&gt;In this section, we will look into the Python API to access the models using&amp;nbsp; &lt;a href=&quot;https://github.com/nomic-ai/pygpt4all&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot;&gt;nomic-ai/pygpt4all&lt;/a&gt;.&lt;/p&gt;
&lt;p&gt;&amp;nbsp;&lt;/p&gt;
&lt;ol&gt;
&lt;li&gt;Install the Python GPT4ALL library using PIP.&lt;/li&gt;
&lt;/ol&gt;
&lt;p&gt;&amp;nbsp;&lt;/p&gt;
&lt;div style=&quot;width: 98%; overflow: auto; padding-left: 10px; padding-bottom: 10px; padding-top: 10px; background: #F5F5F5;&quot;&gt;
&lt;pre&gt;&lt;code&gt;pip install pygpt4all&lt;/code&gt;&lt;/pre&gt;
&lt;/div&gt;
&lt;p&gt;&amp;nbsp;&lt;/p&gt;
&lt;p&gt;&amp;nbsp;&lt;/p&gt;
&lt;p&gt;&amp;nbsp;&lt;/p&gt;
&lt;ol start=&quot;2&quot;&gt;
&lt;li&gt;Download a GPT4All model from &lt;a href=&quot;http://gpt4all.io/models/ggml-gpt4all-l13b-snoozy.bin&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot;&gt;http://gpt4all.io/models/ggml-gpt4all-l13b-snoozy.bin&lt;/a&gt;. You can also browse other models &lt;a href=&quot;https://github.com/nomic-ai/gpt4all-chat#manual-download-of-models&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot;&gt;here&lt;/a&gt;.&lt;/li&gt;
&lt;li&gt;Create a text callback function, load the model, and provide a prompt to &lt;code style=&quot;background: #F5F5F5;&quot;&gt;mode.generate()&lt;/code&gt; function to generate text. Check out the library &lt;a href=&quot;https://nomic-ai.github.io/pygpt4all/&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot;&gt;documentation&lt;/a&gt; to learn more.&lt;/li&gt;
&lt;/ol&gt;
&lt;p&gt;&amp;nbsp;&lt;/p&gt;
&lt;div style=&quot;width: 98%; overflow: auto; padding-left: 10px; padding-bottom: 10px; padding-top: 10px; background: #F5F5F5;&quot;&gt;
&lt;pre&gt;&lt;code&gt;from pygpt4all.models.gpt4all import GPT4All

def new_text_callback(text):
    print(text, end=&quot;&quot;)

model = GPT4All(&quot;./models/ggml-gpt4all-l13b-snoozy.bin&quot;)
model.generate(&quot;Once upon a time, &quot;, n_predict=55, new_text_callback=new_text_callback)
&lt;/code&gt;&lt;/pre&gt;
&lt;/div&gt;
&lt;p&gt;&amp;nbsp;&lt;/p&gt;
&lt;p&gt;&amp;nbsp;&lt;/p&gt;
&lt;p&gt;&amp;nbsp;&lt;/p&gt;
&lt;p&gt;Moreover, you can download and run inference using transformers. Just provide the model name and the version. In our case, we are accessing the latest and improved v1.3-groovy model.&lt;/p&gt;
&lt;p&gt;&amp;nbsp;&lt;/p&gt;
&lt;div style=&quot;width: 98%; overflow: auto; padding-left: 10px; padding-bottom: 10px; padding-top: 10px; background: #F5F5F5;&quot;&gt;
&lt;pre&gt;&lt;code&gt;from transformers import AutoModelForCausalLM

model = AutoModelForCausalLM.from_pretrained(
    &quot;nomic-ai/gpt4all-j&quot;, revision=&quot;v1.3-groovy&quot;
)
&lt;/code&gt;&lt;/pre&gt;
&lt;/div&gt;
&lt;p&gt;&amp;nbsp;&lt;/p&gt;
&lt;p&gt;&amp;nbsp;&lt;/p&gt;
&lt;p&gt;&amp;nbsp;&lt;/p&gt;
&lt;h1&gt;Getting Started&lt;/h1&gt;
&lt;p&gt;&amp;nbsp;&lt;/p&gt;
&lt;p&gt;&amp;nbsp;&lt;/p&gt;
&lt;p&gt;&amp;nbsp;&lt;/p&gt;
&lt;p&gt;The &lt;a href=&quot;https://github.com/nomic-ai/gpt4all&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot;&gt;nomic-ai/gpt4all&lt;/a&gt; repository comes with source code for training and inference, model weights, dataset, and documentation. You can start by trying a few models on your own and then try to integrate it using a Python client or LangChain.&amp;nbsp;&lt;/p&gt;
&lt;p&gt;&amp;nbsp;&lt;/p&gt;
&lt;p&gt;The GPT4ALL provides us with a CPU quantized GPT4All model checkpoint. To access it, we have to:&lt;/p&gt;
&lt;p&gt;&amp;nbsp;&lt;/p&gt;
&lt;ul&gt;
&lt;li&gt;Download the gpt4all-lora-quantized.bin file from &lt;a href=&quot;https://the-eye.eu/public/AI/models/nomic-ai/gpt4all/gpt4all-lora-quantized.bin&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot;&gt;Direct Link&lt;/a&gt; or &lt;a href=&quot;https://tinyurl.com/gpt4all-lora-quantized&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot;&gt;[Torrent-Magnet]&lt;/a&gt;.&lt;/li&gt;
&lt;li&gt;Clone this repository and move the downloaded bin file to &lt;code style=&quot;background: #F5F5F5;&quot;&gt;chat&lt;/code&gt; folder.&amp;nbsp;&lt;/li&gt;
&lt;li&gt;Run the appropriate command to access the model:
&lt;ul&gt;
&lt;li&gt;&lt;strong&gt;M1 Mac/OSX&lt;/strong&gt;: &lt;code style=&quot;background: #F5F5F5;&quot;&gt;cd chat;./gpt4all-lora-quantized-OSX-m1&lt;/code&gt;&lt;/li&gt;
&lt;li&gt;&lt;strong&gt;Linux&lt;/strong&gt;: &lt;code style=&quot;background: #F5F5F5;&quot;&gt;cd chat;./gpt4all-lora-quantized-linux-x86&lt;/code&gt;&lt;/li&gt;
&lt;li&gt;&lt;strong&gt;Windows (PowerShell&lt;/strong&gt;): &lt;code style=&quot;background: #F5F5F5;&quot;&gt;cd chat;./gpt4all-lora-quantized-win64.exe&lt;/code&gt;&lt;/li&gt;
&lt;li&gt;&lt;strong&gt;Intel Mac/OSX&lt;/strong&gt;: &lt;code style=&quot;background: #F5F5F5;&quot;&gt;cd chat;./gpt4all-lora-quantized-OSX-intel&lt;/code&gt;&lt;/li&gt;
&lt;/ul&gt;
&lt;/li&gt;
&lt;/ul&gt;
&lt;p&gt;You can also head to Hugging Face Spaces and try out the &lt;a href=&quot;https://huggingface.co/spaces/Monster/GPT4ALL&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot;&gt;Gpt4all&lt;/a&gt; demo. It is not official, but it is a start.&lt;/p&gt;
&lt;p&gt;&amp;nbsp;&lt;/p&gt;
&lt;p&gt;&amp;nbsp;&lt;/p&gt;
&lt;p&gt;&amp;nbsp;&lt;/p&gt;

Note

Voto oggetto (1-10)

Precisione (1-10)

Tag (5)

✎ Modifica tag 🤖 Suggerisci AI

          dev,nn,llm:20,gpt        

dev 5 gpt 5 llm 5 nn 5 wpg 5

ID: 474 • Tipo: WPG • Reg: 2023-10-09 • Mod: 29/07/2026 20:16