version update: support for interactive mode

README.md

@@ -2,53 +2,80 @@
 Official Python CPU inference for [GPT4All](https://github.com/nomic-ai/gpt4all) language models based on [llama.cpp](https://github.com/ggerganov/llama.cpp) and [ggml](https://github.com/ggerganov/ggml)
 
 [![License: MIT](https://img.shields.io/badge/license-MIT-blue.svg)](https://opensource.org/licenses/MIT)
-
-[//]: # ([![PyPi version](https://badgen.net/pypi/v/ptgpt4all)](https://pypi.org/project/pygpt4all/))
-
-**NB: Under active development**
-
+[![PyPi version](https://badgen.net/pypi/v/pygpt4all)](https://pypi.org/project/pygpt4all/)
+
+<!-- TOC -->
+* [Installation](#installation)
+* [Tutorial](#tutorial)
+    * [Model instantiation](#model-instantiation)
+    * [Simple generation](#simple-generation)
+    * [Interactive Dialogue](#interactive-dialogue)
+* [API reference](#api-reference)
+* [License](#license)
+<!-- TOC -->
 # Installation
 
 ```bash
 pip install pygpt4all
 ```
 
-# Usage
+# Tutorial
+
+You will need first to download the model weights
 
-### GPT4All model
+| Model     | Download link                                            |
+|-----------|----------------------------------------------------------|
+| GPT4ALL   | http://gpt4all.io/models/ggml-gpt4all-l13b-snoozy.bin    |
+| GPT4ALL-j | https://gpt4all.io/models/ggml-gpt4all-j-v1.3-groovy.bin |                                                                     
 
-Download a GPT4All model from http://gpt4all.io/models/ggml-gpt4all-l13b-snoozy.bin
+### Model instantiation
+Once the weights are downloaded, you can instantiate the models as follows:
+* GPT4All model
 
 ```python
-from pygpt4all.models.gpt4all import GPT4All
+from pygpt4all import GPT4All
 
-def new_text_callback(text):
-    print(text, end="")
+model = GPT4All('path/to/ggml-gpt4all-l13b-snoozy.bin')
+```
+
+* GPT4All-J model
 
-model = GPT4All('./models/ggml-gpt4all-l13b-snoozy.bin')
-model.generate("Once upon a time, ", n_predict=55, new_text_callback=new_text_callback)
+```python
+from pygpt4all import GPT4All_J
+
+model = GPT4All_J('path/to/ggml-gpt4all-j-v1.3-groovy.bin')
 ```
 
-### GPT4All-J model
 
-Download the GPT4All-J model from https://gpt4all.io/models/ggml-gpt4all-j-v1.3-groovy.bin
+### Simple generation
+The `generate` function is used to generate new tokens from the `prompt` given as input:
 
 ```python
-from pygpt4all.models.gpt4all_j import GPT4All_J
+for token in model.generate("Tell me a joke ?\n"):
+    print(token, end='', flush=True)
+```
 
-def new_text_callback(text):
-    print(text, end="")
+### Interactive Dialogue
+You can set up an interactive dialogue by simply keeping the `model` variable alive:
 
-model = GPT4All_J('./models/ggml-gpt4all-j-v1.3-groovy.bin')
-model.generate("Once upon a time, ", n_predict=55, new_text_callback=new_text_callback)
+```python
+while True:
+    try:
+        prompt = input("You: ", flush=True)
+        if prompt == '':
+            continue
+        print(f"AI:", end='')
+        for token in model.generate(prompt):
+            print(f"{token}", end='', flush=True)
+        print()
+    except KeyboardInterrupt:
+        break
 ```
 
-[//]: # (* You can always refer to the [short documentation]&#40;https://nomic-ai.github.io/pyllamacpp/&#41; for more details.)
+# API reference
+You can check the [API reference documentation](https://nomic-ai.github.io/pygpt4all/) for more details.
 
 
 # License
-
-
-
 This project is licensed under the MIT  [License](./LICENSE).
 

pygpt4all/__init__.py

@@ -0,0 +1,2 @@
+from pygpt4all.models.gpt4all import GPT4All
+from pygpt4all.models.gpt4all_j import GPT4All_J

pygpt4all/models/gpt4all.py

@@ -2,7 +2,7 @@
 # -*- coding: utf-8 -*-
 
 """
-GPT4ALL with `llama.cpp` backend
+GPT4ALL with `llama.cpp` backend through [pyllamacpp](https://github.com/abdeladim-s/pyllamacpp)
 """
 
 __author__ = "abdeladim-s"
@@ -24,47 +24,54 @@ class GPT4All(pyllamacpp.model.Model):
     ```python
     from pygpt4all.models.gpt4all import GPT4All
 
-    def new_text_callback(text):
-        print(text, end="")
-
-    model = GPT4All('./models/ggml-gpt4all-j.bin')
-    model.generate("Once upon a time, ", n_predict=55, new_text_callback=new_text_callback)
+    model = GPT4All('path/to/gpt4all/model')
+    for token in model.generate("Tell me a joke ?"):
+        print(token, end='', flush=True)
     ```
     """
 
     def __init__(self,
                  model_path: str,
+                 prompt_context: str = '',
+                 prompt_prefix: str = '',
+                 prompt_suffix: str = '',
+                 log_level: int = logging.ERROR,
                  n_ctx: int = 512,
-                 n_parts: int = -1,
                  seed: int = 0,
+                 n_parts: int = -1,
                  f16_kv: bool = False,
                  logits_all: bool = False,
                  vocab_only: bool = False,
                  use_mlock: bool = False,
-                 embedding: bool = False,
-                 log_level: int = logging.INFO):
+                 embedding: bool = False):
         """
-        :param model_path: The path to a gpt4all `ggml` model
-        :param n_ctx: context size
-        :param n_parts:
-        :param seed: RNG seed, 0 for random
+        :param model_path: the path to the gpt4all model
+        :param prompt_context: the global context of the interaction
+        :param prompt_prefix: the prompt prefix
+        :param prompt_suffix: the prompt suffix
+        :param log_level: logging level, set to INFO by default
+        :param n_ctx: LLaMA context
+        :param seed: random seed
+        :param n_parts: LLaMA n_parts
         :param f16_kv: use fp16 for KV cache
         :param logits_all: the llama_eval() call computes all logits, not just the last one
         :param vocab_only: only load the vocabulary, no weights
         :param use_mlock: force system to keep model in RAM
         :param embedding: embedding mode only
-        :param log_level: logging level, set to INFO by default
         """
         # set logging level
         set_log_level(log_level)
-        super(GPT4All, self).__init__(ggml_model=model_path,
+        super(GPT4All, self).__init__(model_path=model_path,
+                                      prompt_context=prompt_context,
+                                      prompt_prefix=prompt_prefix,
+                                      prompt_suffix=prompt_suffix,
+                                      log_level=log_level,
                                       n_ctx=n_ctx,
-                                      n_parts=n_parts,
                                       seed=seed,
+                                      n_parts=n_parts,
                                       f16_kv=f16_kv,
                                       logits_all=logits_all,
                                       vocab_only=vocab_only,
                                       use_mlock=use_mlock,
-                                      embedding=embedding,
-                                      log_level=log_level)
+                                      embedding=embedding)
 

pygpt4all/models/gpt4all_j.py

@@ -23,21 +23,29 @@ class GPT4All_J(pygptj.model.Model):
     ```python
     from pygpt4all.models.gpt4all_j import GPT4All_J
 
-    def new_text_callback(text):
-        print(text, end="")
-
-    model = GPT4All_J('./models/ggml-gpt4all-j.bin')
-    model.generate("Once upon a time, ", n_predict=55, new_text_callback=new_text_callback)
+    model = GPT4All_J('.path/to/gpr4all-j/model')
+    for token in model.generate("Tell me a joke ?"):
+        print(token, end='', flush=True)
     ```
     """
 
     def __init__(self,
                  model_path: str,
-                 log_level: int = logging.INFO):
+                 prompt_context: str = '',
+                 prompt_prefix: str = '',
+                 prompt_suffix: str = '',
+                 log_level: int = logging.ERROR):
         """
-        :param model_path: The path to a gpt4all `ggml` model
-        :param log_level: logging level, set to INFO by default
+        :param model_path: The path to a gpt4all-j model
+        :param prompt_context: the global context of the interaction
+        :param prompt_prefix: the prompt prefix
+        :param prompt_suffix: the prompt suffix
+        :param log_level: logging level
         """
         # set logging level
         set_log_level(log_level)
-        super(GPT4All_J, self).__init__(model_path=model_path, log_level=log_level)
+        super(GPT4All_J, self).__init__(model_path=model_path,
+                                        prompt_context=prompt_context,
+                                        prompt_prefix=prompt_prefix,
+                                        prompt_suffix=prompt_suffix,
+                                        log_level=log_level)

setup.py

@@ -10,7 +10,7 @@
 
 setup(
     name="pygpt4all",
-    version="1.0.1",
+    version="1.1.0",
     author="Abdeladim Sadiki",
     description="Official Python CPU inference for GPT4All language models based on llama.cpp and ggml",
     long_description=long_description,
@@ -21,11 +21,10 @@
     package_dir={'': '.'},
     long_description_content_type="text/markdown",
     license='MIT',
-    # project_urls={
-    #     'Documentation': '',
-    #     'Source': '',
-    #     'Tracker': '',
-    # },
-    install_requires=["pyllamacpp==1.0.7", "pygptj"],
-    # extras_require={"all": [""]},
+    project_urls={
+        'Documentation': 'nomic-ai.github.io/pygpt4all/',
+        'Source': 'https://github.com/nomic-ai/pygpt4all',
+        'Tracker': 'https://github.com/nomic-ai/pygpt4all/issues',
+    },
+    install_requires=["pyllamacpp", "pygptj"],
 )