Merge pull request #134 from MervinPraison/develop

Develop

Author: MervinPraison

.praisonignore

@@ -2,5 +2,4 @@ cookbooks
 tests
 docs
 praisonai/ui/*
-praisonai/inbuilt_tools/*
-praisonai/inc/*
+praisonai/inbuilt_tools/*

Dockerfile

@@ -1,6 +1,6 @@
 FROM python:3.11-slim
 WORKDIR /app
 COPY . .
-RUN pip install flask praisonai==0.0.59rc9 gunicorn markdown
+RUN pip install flask praisonai==0.0.59rc11 gunicorn markdown
 EXPOSE 8080
 CMD ["gunicorn", "-b", "0.0.0.0:8080", "api:app"]

config.yaml

@@ -1,5 +1,29 @@
+ollama_save: "false"
+huggingface_save: "false"
+train: "false"
+
 model_name: "unsloth/Meta-Llama-3.1-8B-Instruct-bnb-4bit"
 hf_model_name: "mervinpraison/llama-3.1-tamilan-8B-test"
+ollama_model: "mervinpraison/llama3.1-tamilan-test"
+model_parameters: "8b"
+
+dataset:
+  - name: "yahma/alpaca-cleaned"
+    split_type: "train"
+    processing_func: "format_prompts"
+    rename:
+      input: "input"
+      output: "output"
+      instruction: "instruction"
+    filter_data: false
+    filter_column_value: "id"
+    filter_value: "alpaca"
+    num_samples: 20000
+
+dataset_text_field: "text"
+dataset_num_proc: 2
+packing: false
+
 max_seq_length: 2048
 load_in_4bit: true
 lora_r: 16
@@ -19,23 +43,6 @@ random_state: 3407
 use_rslora: false
 loftq_config: null
 
-dataset:
-  - name: "yahma/alpaca-cleaned"
-    split_type: "train"
-    processing_func: "format_prompts"
-    rename:
-      input: "input"
-      output: "output"
-      instruction: "instruction"
-    filter_data: false
-    filter_column_value: "id"
-    filter_value: "alpaca"
-    num_samples: 20000
-
-dataset_text_field: "text"
-dataset_num_proc: 2
-packing: false
-
 per_device_train_batch_size: 2
 gradient_accumulation_steps: 2
 warmup_steps: 5
@@ -50,7 +57,4 @@ seed: 3407
 output_dir: "outputs"
 
 quantization_method: 
-  - "q4_k_m"
-
-ollama_model: "llama3.1-tamilan-test"
-model_parameters: "8b"
+  - "q4_k_m"

docs/api/praisonai/deploy.html

@@ -110,7 +110,7 @@ <h2 id="raises">Raises</h2>
             file.write(&#34;FROM python:3.11-slim\n&#34;)
             file.write(&#34;WORKDIR /app\n&#34;)
             file.write(&#34;COPY . .\n&#34;)
-            file.write(&#34;RUN pip install flask praisonai==0.0.59rc9 gunicorn markdown\n&#34;)
+            file.write(&#34;RUN pip install flask praisonai==0.0.59rc11 gunicorn markdown\n&#34;)
             file.write(&#34;EXPOSE 8080\n&#34;)
             file.write(&#39;CMD [&#34;gunicorn&#34;, &#34;-b&#34;, &#34;0.0.0.0:8080&#34;, &#34;api:app&#34;]\n&#39;)
 

praisonai.rb

@@ -3,7 +3,7 @@ class Praisonai < Formula
 
     desc "AI tools for various AI applications"
     homepage "https://github.com/MervinPraison/PraisonAI"
-    url "https://github.com/MervinPraison/PraisonAI/archive/refs/tags/0.0.59rc9.tar.gz"
+    url "https://github.com/MervinPraison/PraisonAI/archive/refs/tags/0.0.59rc11.tar.gz"
     sha256 "1828fb9227d10f991522c3f24f061943a254b667196b40b1a3e4a54a8d30ce32"  # Replace with actual SHA256 checksum
     license "MIT"
 

praisonai/cli.py

@@ -12,6 +12,7 @@
 from .auto import AutoGenerator
 from .agents_generator import AgentsGenerator
 from .inbuilt_tools import *
+from .inc.config import generate_config
 import shutil
 import subprocess
 import logging
@@ -130,17 +131,20 @@ def main(self):
 
         if args.agent_file == 'train':
             package_root = os.path.dirname(os.path.abspath(__file__))
-            config_yaml_source = os.path.join(package_root, 'setup', 'config.yaml')
             config_yaml_destination = os.path.join(os.getcwd(), 'config.yaml')
 
-            if not os.path.exists(config_yaml_destination):
-                try:
-                    shutil.copyfile(config_yaml_source, config_yaml_destination)
-                    print("config.yaml copied to the current directory.")
-                except FileExistsError:
-                    print("config.yaml already exists in the current directory. Skipping copy.")
-            else:
-                print("config.yaml already exists in the current directory. Skipping copy.")
+            # Generate config.yaml using the function
+            config = generate_config(
+                model_name=args.model,
+                hf_model_name=args.hf,
+                ollama_model_name=args.ollama,
+                dataset=[{
+                    "name": args.dataset
+                }]
+            )
+            with open('config.yaml', 'w') as f:
+                yaml.dump(config, f, default_flow_style=False, indent=2) 
+
 
             if 'init' in sys.argv:
                 from praisonai.setup.setup_conda_env import main as setup_conda_main
@@ -150,12 +154,12 @@ def main(self):
 
             try:
                 result = subprocess.check_output(['conda', 'env', 'list'])
-                if 'prasion_env' in result.decode('utf-8'):
-                    print("Conda environment 'prasion_env' found.")
+                if 'praison_env' in result.decode('utf-8'):
+                    print("Conda environment 'praison_env' found.")
                 else:
                     raise subprocess.CalledProcessError(1, 'grep')
             except subprocess.CalledProcessError:
-                print("Conda environment 'prasion_env' not found. Setting it up...")
+                print("Conda environment 'praison_env' not found. Setting it up...")
                 from praisonai.setup.setup_conda_env import main as setup_conda_main
                 setup_conda_main()
                 print("All packages installed.")
@@ -167,7 +171,7 @@ def main(self):
             env = os.environ.copy()
             env['PYTHONUNBUFFERED'] = '1'
 
-            stream_subprocess(['conda', 'run', '--no-capture-output', '--name', 'prasion_env', 'python', '-u', train_script_path, 'train'] + train_args, env=env)
+            stream_subprocess(['conda', 'run', '--no-capture-output', '--name', 'praison_env', 'python', '-u', train_script_path, 'train'] + train_args, env=env)
             return
 
         invocation_cmd = "praisonai"
@@ -240,7 +244,11 @@ def parse_args(self):
         parser.add_argument("--auto", nargs=argparse.REMAINDER, help="Enable auto mode and pass arguments for it")
         parser.add_argument("--init", nargs=argparse.REMAINDER, help="Enable auto mode and pass arguments for it")
         parser.add_argument("agent_file", nargs="?", help="Specify the agent file")
-        parser.add_argument("--deploy", action="store_true", help="Deploy the application")  # New argument
+        parser.add_argument("--deploy", action="store_true", help="Deploy the application") 
+        parser.add_argument("--model", type=str, help="Model name")
+        parser.add_argument("--hf", type=str, help="Hugging Face model name")
+        parser.add_argument("--ollama", type=str, help="Ollama model name")
+        parser.add_argument("--dataset", type=str, help="Dataset name for training", default="yahma/alpaca-cleaned")
         args, unknown_args = parser.parse_known_args()
 
         if unknown_args and unknown_args[0] == '-b' and unknown_args[1] == 'api:app':

praisonai/deploy.py

@@ -56,7 +56,7 @@ def create_dockerfile(self):
             file.write("FROM python:3.11-slim\n")
             file.write("WORKDIR /app\n")
             file.write("COPY . .\n")
-            file.write("RUN pip install flask praisonai==0.0.59rc9 gunicorn markdown\n")
+            file.write("RUN pip install flask praisonai==0.0.59rc11 gunicorn markdown\n")
             file.write("EXPOSE 8080\n")
             file.write('CMD ["gunicorn", "-b", "0.0.0.0:8080", "api:app"]\n')
 

praisonai/inc/config.py

@@ -0,0 +1,96 @@
+def generate_config(
+    ollama_save=None,
+    huggingface_save=None,
+    train=None,
+    model_name=None,
+    hf_model_name=None,
+    ollama_model_name=None,
+    model_parameters=None,
+    max_seq_length=None,
+    load_in_4bit=None,
+    lora_r=None,
+    lora_target_modules=None,
+    lora_alpha=None,
+    lora_dropout=None,
+    lora_bias=None,
+    use_gradient_checkpointing=None,
+    random_state=None,
+    use_rslora=None,
+    loftq_config=None,
+    dataset=None,
+    dataset_text_field=None,
+    dataset_num_proc=None,
+    packing=None,
+    per_device_train_batch_size=None,
+    gradient_accumulation_steps=None,
+    warmup_steps=None,
+    num_train_epochs=None,
+    max_steps=None,
+    learning_rate=None,
+    logging_steps=None,
+    optim=None,
+    weight_decay=None,
+    lr_scheduler_type=None,
+    seed=None,
+    output_dir=None,
+    quantization_method=None
+):
+    """Generates the configuration for PraisonAI with dynamic overrides."""
+
+    config = {
+        "ollama_save": ollama_save or "false",
+        "huggingface_save": huggingface_save or "false",
+        "train": train or "false",
+
+        "model_name": model_name or "unsloth/Meta-Llama-3.1-8B-Instruct-bnb-4bit",
+        "hf_model_name": hf_model_name or "mervinpraison/llama-3.1-tamilan-8B-test",
+        "ollama_model": ollama_model_name or "mervinpraison/llama3.1-tamilan-test",
+        "model_parameters": model_parameters or "8b",
+
+        "dataset": dataset or [
+            {
+                "name": "yahma/alpaca-cleaned",
+                "split_type": "train",
+                "processing_func": "format_prompts",
+                "rename": {"input": "input", "output": "output", "instruction": "instruction"},
+                "filter_data": False,
+                "filter_column_value": "id",
+                "filter_value": "alpaca",
+                "num_samples": 20000
+            }
+        ],
+
+        "dataset_text_field": dataset_text_field or "text",
+        "dataset_num_proc": dataset_num_proc or 2,
+        "packing": packing or False,
+
+        "max_seq_length": max_seq_length or 2048,
+        "load_in_4bit": load_in_4bit or True,
+        "lora_r": lora_r or 16,
+        "lora_target_modules": lora_target_modules or [
+            "q_proj", "k_proj", "v_proj", "o_proj", "gate_proj", "up_proj", "down_proj"
+        ],
+        "lora_alpha": lora_alpha or 16,
+        "lora_dropout": lora_dropout or 0,
+        "lora_bias": lora_bias or "none",
+        "use_gradient_checkpointing": use_gradient_checkpointing or "unsloth",
+        "random_state": random_state or 3407,
+        "use_rslora": use_rslora or False,
+        "loftq_config": loftq_config or None,
+
+        "per_device_train_batch_size": per_device_train_batch_size or 2,
+        "gradient_accumulation_steps": gradient_accumulation_steps or 2,
+        "warmup_steps": warmup_steps or 5,
+        "num_train_epochs": num_train_epochs or 1,
+        "max_steps": max_steps or 10,
+        "learning_rate": learning_rate or 2.0e-4,
+        "logging_steps": logging_steps or 1,
+        "optim": optim or "adamw_8bit",
+        "weight_decay": weight_decay or 0.01,
+        "lr_scheduler_type": lr_scheduler_type or "linear",
+        "seed": seed or 3407,
+        "output_dir": output_dir or "outputs",
+
+        "quantization_method": quantization_method or ["q4_k_m"]
+    }
+    return config

praisonai/setup/config.yaml

@@ -1,5 +1,12 @@
+ollama_save: "true"
+huggingface_save: "true"
+train: "true"
+
 model_name: "unsloth/Meta-Llama-3.1-8B-Instruct-bnb-4bit"
 hf_model_name: "mervinpraison/llama-3.1-tamilan-8B-test"
+ollama_model: "mervinpraison/llama3.1-tamilan-test"
+model_parameters: "8b"
+
 max_seq_length: 2048
 load_in_4bit: true
 lora_r: 16
@@ -50,7 +57,4 @@ seed: 3407
 output_dir: "outputs"
 
 quantization_method: 
-  - "q4_k_m"
-
-ollama_model: "llama3.1-tamilan-test"
-model_parameters: "8b"
+  - "q4_k_m"

praisonai/setup/setup_conda_env.sh

@@ -33,7 +33,7 @@ else
 fi
 
 # Create and activate the Conda environment
-ENV_NAME="prasion_env"
+ENV_NAME="praison_env"
 if conda info --envs | grep -q $ENV_NAME; then
     echo "Environment $ENV_NAME already exists. Recreating..."
     conda env remove -y -n $ENV_NAME  # Remove existing environment