Skip to content

Overview

Initialization

  1. model_name (string) ("EricFillion/smollm3-3b-mlx"). Provide a string that contains a Hugging Face ID or a path to a model directory. We recommend you provide a model that has already converted to MLX.
from erictransformer import EricChatMLX

eric_chat = EricChatMLX(model_name="EricFillion/smollm3-3b-mlx")
Recommended Models
EricFillion/smollm3-3b-mlx
EricFillion/gpt-oss-20b-mlx
EricFillion/gpt-oss-120b-mlx

We recommended you use either an OpenAI GPT-OSS model or a Hugging Face SmolLM3-3B as we support their chat templates.

Call

Arguments:

  1. text (string or List[dict]): Either an input user prompt or a list of dictionaries that contain a conversation history

  2. args (CHATCallArgs) (CHATCallArgs()): See this webpage for more detail.

from erictransformer import EricChatMLX, CHATCallArgs

eric_chat = EricChatMLX(model_name="EricFillion/smollm3-3b-mlx")

args = CHATCallArgs(  # Min/max number of tokens to produce during generation.
    min_len=1,
    max_len=1024,
    # Sampling settings. 
    temp=1.0,
    top_k=32,
    top_p=0.6)

result = eric_chat("Hello world", args=args)

print(result.text)  # str 

# Conversation 
convo = [{"role": "system", "content": "You are a friendly ai assistant"},
         {"role": "user", "content": "what's 1 +1?"},
         {"role": "assistant", "content": "2"},
         {"role": "user", "content": "what's 2 +2?"}
         ]

result = eric_chat(convo, args=args)

print(result)  # str

stream()

The stream() method contains 2 arguments:

  1. text (string or List[dict]): Either an input user prompt or a list of dictionaries that contain a conversation history.

  2. args (CHATCallArgs) (CHATCallArgs()): See this webpage for more detail.

from erictransformer import EricChatMLX, CHATCallArgs

eric_chat = EricChatMLX(model_name="EricFillion/smollm3-3b-mlx")

args = CHATCallArgs(  # Min/max number of tokens to produce during generation.
    min_len=1,
    max_len=1024,
    # Sampling settings.
    temp=1.0,
    top_k=50,
    top_p=1.0)

text = "Hello world"  # you can instead provide a conversation like in the previous example

for token in eric_chat.stream(text, args=args):
    if token.marker == "text":
        print(token.text, end="")
    # you can also see the thinking tokens with token.marker == "thinking"

RAG: EricSearch() + EricChatMLX()

EricSearch can be integrated with the EricChatMLX in the same way its be integrated with EricChat. See the bottom of this page.