Instructions to use canxp-ai/canxpai-maplept-mini-v1 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use canxp-ai/canxpai-maplept-mini-v1 with Transformers:

# Use a pipeline as a high-level helper
from transformers import pipeline

pipe = pipeline("text-generation", model="canxp-ai/canxpai-maplept-mini-v1", trust_remote_code=True)
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# pip install -U transformers accelerate
# Load model directly
from transformers import AutoModel
model = AutoModel.from_pretrained("canxp-ai/canxpai-maplept-mini-v1", trust_remote_code=True, device_map="auto")

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use canxp-ai/canxpai-maplept-mini-v1 with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "canxp-ai/canxpai-maplept-mini-v1"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "canxp-ai/canxpai-maplept-mini-v1",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/canxp-ai/canxpai-maplept-mini-v1

SGLang

How to use canxp-ai/canxpai-maplept-mini-v1 with SGLang:

Install from pip and serve model

# Install SGLang from pip:
pip install sglang
# Start the SGLang server:
python3 -m sglang.launch_server \
    --model-path "canxp-ai/canxpai-maplept-mini-v1" \
    --host 0.0.0.0 \
    --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "canxp-ai/canxpai-maplept-mini-v1",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker images

docker run --gpus all \
    --shm-size 32g \
    -p 30000:30000 \
    -v ~/.cache/huggingface:/root/.cache/huggingface \
    --env "HF_TOKEN=<secret>" \
    --ipc=host \
    lmsysorg/sglang:latest \
    python3 -m sglang.launch_server \
        --model-path "canxp-ai/canxpai-maplept-mini-v1" \
        --host 0.0.0.0 \
        --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "canxp-ai/canxpai-maplept-mini-v1",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use canxp-ai/canxpai-maplept-mini-v1 with Docker Model Runner:
```
docker model run hf.co/canxp-ai/canxpai-maplept-mini-v1
```

canxpai-maplept-mini-v1 / modeling_maplept.py

vamman

Update model weights/config

100ae20 verified about 1 year ago

Raw History Blame Contribute Delete

1.81 kB

from transformers.models.phi3.modeling_phi3 import Phi3ForCausalLM
from transformers import AutoConfig, AutoModelForCausalLM
from .configuration_maplept import MaplePTConfig
import torch.nn as nn


class MaplePTForCausalLM(Phi3ForCausalLM):
    config_class = MaplePTConfig
    model_type = "maplept"

    def __init__(self, config):
        if not isinstance(config, MaplePTConfig):
            config = MaplePTConfig.from_dict(config.to_dict())
        super().__init__(config)

        base = getattr(self, "model", None)
        if base is None:
            return

        # --- ensure embed_tokens alias ---
        if hasattr(base, "embed_tokens"):
            self.embed_tokens = base.embed_tokens

        # --- ensure model.layers alias (vLLM critical) ---
        # vLLM searches model.layers.*, but Phi3 nests them in model.model.layers
        if not hasattr(base, "layers"):
            try:
                base.layers = base.model.layers
                print("✅ Aliased base.layers → base.model.layers")
            except AttributeError:
                print("⚠️ Could not alias layers (missing model.model.layers)")

        # --- ensure top-level direct alias too ---
        if not hasattr(self, "layers") and hasattr(base, "layers"):
            self.layers = base.layers
            print("✅ Aliased self.layers → model.layers")

        # --- ensure normalization aliases (for completeness) ---
        if not hasattr(base, "norm"):
            if hasattr(base, "final_layernorm"):
                base.norm = base.final_layernorm
            elif hasattr(base, "model") and hasattr(base.model, "final_layernorm"):
                base.norm = base.model.final_layernorm
        if not hasattr(self, "norm") and hasattr(base, "norm"):
            self.norm = base.norm

	from transformers.models.phi3.modeling_phi3 import Phi3ForCausalLM
	from transformers import AutoConfig, AutoModelForCausalLM
	from .configuration_maplept import MaplePTConfig
	import torch.nn as nn


	class MaplePTForCausalLM(Phi3ForCausalLM):
	config_class = MaplePTConfig
	model_type = "maplept"

	def __init__(self, config):
	if not isinstance(config, MaplePTConfig):
	config = MaplePTConfig.from_dict(config.to_dict())
	super().__init__(config)

	base = getattr(self, "model", None)
	if base is None:
	return

	# --- ensure embed_tokens alias ---
	if hasattr(base, "embed_tokens"):
	self.embed_tokens = base.embed_tokens

	# --- ensure model.layers alias (vLLM critical) ---
	# vLLM searches model.layers.*, but Phi3 nests them in model.model.layers
	if not hasattr(base, "layers"):
	try:
	base.layers = base.model.layers
	print("✅ Aliased base.layers → base.model.layers")
	except AttributeError:
	print("⚠️ Could not alias layers (missing model.model.layers)")

	# --- ensure top-level direct alias too ---
	if not hasattr(self, "layers") and hasattr(base, "layers"):
	self.layers = base.layers
	print("✅ Aliased self.layers → model.layers")

	# --- ensure normalization aliases (for completeness) ---
	if not hasattr(base, "norm"):
	if hasattr(base, "final_layernorm"):
	base.norm = base.final_layernorm
	elif hasattr(base, "model") and hasattr(base.model, "final_layernorm"):
	base.norm = base.model.final_layernorm
	if not hasattr(self, "norm") and hasattr(base, "norm"):
	self.norm = base.norm