Instructions to use Jashan887/97_Learning_Unit_L1 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use Jashan887/97_Learning_Unit_L1 with Transformers:

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

pipe = pipeline("text-generation", model="Jashan887/97_Learning_Unit_L1", trust_remote_code=True)
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoModelForCausalLM
model = AutoModelForCausalLM.from_pretrained("Jashan887/97_Learning_Unit_L1", trust_remote_code=True, dtype="auto")

Notebooks
Google Colab
Kaggle
Local Apps

vLLM

How to use Jashan887/97_Learning_Unit_L1 with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "Jashan887/97_Learning_Unit_L1"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "Jashan887/97_Learning_Unit_L1",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/Jashan887/97_Learning_Unit_L1

SGLang

How to use Jashan887/97_Learning_Unit_L1 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 "Jashan887/97_Learning_Unit_L1" \
    --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": "Jashan887/97_Learning_Unit_L1",
		"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 "Jashan887/97_Learning_Unit_L1" \
        --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": "Jashan887/97_Learning_Unit_L1",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use Jashan887/97_Learning_Unit_L1 with Docker Model Runner:
```
docker model run hf.co/Jashan887/97_Learning_Unit_L1
```

97_Learning_Unit_L1 / modeling_gravity_moe.py

Jashan887

Upload folder using huggingface_hub

112d755 verified 25 days ago

raw

history blame contribute delete

2.01 kB

	# Copyright 2026 Trillion Labs and the HuggingFace Inc. team. All rights reserved.

	# Licensed under the Apache License, Version 2.0 (the "License");
	# you may not use this file except in compliance with the License.
	# You may obtain a copy of the License at
	#
	# http://www.apache.org/licenses/LICENSE-2.0
	#
	# Unless required by applicable law or agreed to in writing, software
	# distributed under the License is distributed on an "AS IS" BASIS,
	# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
	# See the License for the specific language governing permissions and
	# limitations under the License.
	"""
	GravityMoE model — inherits from DeepSeek V3.

	GravityMoE shares the same sparse Mixture-of-Experts architecture as DeepSeek V3
	(MLA attention, sigmoid routing with bias correction, shared + routed experts)
	but with different model hyperparameters. All modeling logic is inherited from
	the DeepSeek V3 implementation in `transformers`.
	"""

	from transformers.conversion_mapping import _MODEL_TO_CONVERSION_PATTERN
	from transformers.models.deepseek_v3.modeling_deepseek_v3 import (
	DeepseekV3ForCausalLM,
	DeepseekV3Model,
	DeepseekV3PreTrainedModel,
	)

	from .configuration_gravity_moe import GravityMoEConfig

	# Register weight conversion so that from_pretrained fuses per-expert
	# checkpoint weights (experts.*.gate_proj, etc.) into 3D tensors
	# (experts.gate_up_proj, experts.down_proj), same as DeepSeek V3.
	_MODEL_TO_CONVERSION_PATTERN["gravity_moe"] = "qwen2_moe"


	class GravityMoEPreTrainedModel(DeepseekV3PreTrainedModel):
	config_class = GravityMoEConfig
	_keep_in_fp32_modules_strict = ["e_score_correction_bias"]
	_keys_to_ignore_on_load_unexpected = [r"model\.layers\.28.*"]


	class GravityMoEModel(DeepseekV3Model):
	config_class = GravityMoEConfig


	class GravityMoEForCausalLM(DeepseekV3ForCausalLM):
	config_class = GravityMoEConfig


	__all__ = [
	"GravityMoEPreTrainedModel",
	"GravityMoEModel",
	"GravityMoEForCausalLM",
	]