Text Generation
Transformers
Safetensors
English
pruned_flex_olmo
custom_code
math
pruned
distilled
mixture-of-experts
Instructions to use hbfreed/flex-math-5504 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Transformers
How to use hbfreed/flex-math-5504 with Transformers:
# Use a pipeline as a high-level helper from transformers import pipeline pipe = pipeline("text-generation", model="hbfreed/flex-math-5504", trust_remote_code=True)# Load model directly from transformers import AutoModelForCausalLM model = AutoModelForCausalLM.from_pretrained("hbfreed/flex-math-5504", trust_remote_code=True, dtype="auto") - Notebooks
- Google Colab
- Kaggle
- Local Apps
- vLLM
How to use hbfreed/flex-math-5504 with vLLM:
Install from pip and serve model
# Install vLLM from pip: pip install vllm # Start the vLLM server: vllm serve "hbfreed/flex-math-5504" # Call the server using curl (OpenAI-compatible API): curl -X POST "http://localhost:8000/v1/completions" \ -H "Content-Type: application/json" \ --data '{ "model": "hbfreed/flex-math-5504", "prompt": "Once upon a time,", "max_tokens": 512, "temperature": 0.5 }'Use Docker
docker model run hf.co/hbfreed/flex-math-5504
- SGLang
How to use hbfreed/flex-math-5504 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 "hbfreed/flex-math-5504" \ --host 0.0.0.0 \ --port 30000 # Call the server using curl (OpenAI-compatible API): curl -X POST "http://localhost:30000/v1/completions" \ -H "Content-Type: application/json" \ --data '{ "model": "hbfreed/flex-math-5504", "prompt": "Once upon a time,", "max_tokens": 512, "temperature": 0.5 }'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 "hbfreed/flex-math-5504" \ --host 0.0.0.0 \ --port 30000 # Call the server using curl (OpenAI-compatible API): curl -X POST "http://localhost:30000/v1/completions" \ -H "Content-Type: application/json" \ --data '{ "model": "hbfreed/flex-math-5504", "prompt": "Once upon a time,", "max_tokens": 512, "temperature": 0.5 }' - Docker Model Runner
How to use hbfreed/flex-math-5504 with Docker Model Runner:
docker model run hf.co/hbfreed/flex-math-5504
Upload configuration_pruned_flex_olmo.py with huggingface_hub
Browse files
configuration_pruned_flex_olmo.py
ADDED
|
@@ -0,0 +1,19 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
"""Configuration for pruned FlexOlmo models with variable-width expert 1."""
|
| 2 |
+
|
| 3 |
+
from transformers import FlexOlmoConfig
|
| 4 |
+
|
| 5 |
+
|
| 6 |
+
class PrunedFlexOlmoConfig(FlexOlmoConfig):
|
| 7 |
+
"""Config for FlexOlmo with a pruned expert 1.
|
| 8 |
+
|
| 9 |
+
Extends FlexOlmoConfig with expert_1_intermediate_size to specify
|
| 10 |
+
the width of the pruned expert.
|
| 11 |
+
"""
|
| 12 |
+
|
| 13 |
+
model_type = "pruned_flex_olmo"
|
| 14 |
+
|
| 15 |
+
def __init__(self, expert_1_intermediate_size: int = None, **kwargs):
|
| 16 |
+
super().__init__(**kwargs)
|
| 17 |
+
# expert_1_intermediate_size: width of pruned expert 1
|
| 18 |
+
# If None, falls back to intermediate_size (no pruning)
|
| 19 |
+
self.expert_1_intermediate_size = expert_1_intermediate_size or self.intermediate_size
|