Text Generation
Transformers
Safetensors
English
mume_gpt
causal-lm
custom_code
math
base-model
Eval Results (legacy)
Instructions to use MuseMesh/mume-math-125m with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Transformers
How to use MuseMesh/mume-math-125m with Transformers:
# Use a pipeline as a high-level helper from transformers import pipeline pipe = pipeline("text-generation", model="MuseMesh/mume-math-125m", trust_remote_code=True)# pip install -U transformers accelerate # Load model directly from transformers import AutoModelForCausalLM model = AutoModelForCausalLM.from_pretrained("MuseMesh/mume-math-125m", trust_remote_code=True, device_map="auto") - Notebooks
- Google Colab
- Kaggle
- Local Apps Settings
- vLLM
How to use MuseMesh/mume-math-125m with vLLM:
Install from pip and serve model
# Install vLLM from pip: pip install vllm # Start the vLLM server: vllm serve "MuseMesh/mume-math-125m" # Call the server using curl (OpenAI-compatible API): curl -X POST "http://localhost:8000/v1/completions" \ -H "Content-Type: application/json" \ --data '{ "model": "MuseMesh/mume-math-125m", "prompt": "Once upon a time,", "max_tokens": 512, "temperature": 0.5 }'Use Docker
docker model run hf.co/MuseMesh/mume-math-125m
- SGLang
How to use MuseMesh/mume-math-125m 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 "MuseMesh/mume-math-125m" \ --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": "MuseMesh/mume-math-125m", "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 "MuseMesh/mume-math-125m" \ --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": "MuseMesh/mume-math-125m", "prompt": "Once upon a time,", "max_tokens": 512, "temperature": 0.5 }' - Docker Model Runner
How to use MuseMesh/mume-math-125m with Docker Model Runner:
docker model run hf.co/MuseMesh/mume-math-125m
Download eval/verification.json from MuseMesh/mume-math-125m: direct link, hf CLI and curl.
- Browser
- Download file 5.15 kB
-
https://huggingface.co/MuseMesh/mume-math-125m/resolve/main/eval/verification.json
- Command line
-
hf download hf://MuseMesh/mume-math-125m/eval/verification.json
-
curl -L -o verification.json https://huggingface.co/MuseMesh/mume-math-125m/resolve/main/eval/verification.json
5.15 kB
| { | |
| "repo": "mume-math-125m", | |
| "version": "v0.1.0", | |
| "run": "math0_shared32k_d125m_ctx1024", | |
| "torch": "2.14.0+cu130", | |
| "transformers": "5.18.0", | |
| "hf_load_seconds": 0.3, | |
| "hf_param_counts": { | |
| "total": 134105856, | |
| "embedding": 24576000, | |
| "head": 24576000, | |
| "non_embedding": 84953856 | |
| }, | |
| "tokenizer_class": "MumeTokenizer", | |
| "orig_param_counts": { | |
| "total": 134105856, | |
| "embedding": 24576000, | |
| "token_embedding": 24576000, | |
| "position_embedding": 0, | |
| "value_embedding": 0, | |
| "head": 24576000, | |
| "non_embedding": 84953856 | |
| }, | |
| "hf_head_equals_wte": false, | |
| "logits": [ | |
| { | |
| "input": "owm_val", | |
| "tokens": 619, | |
| "max_abs_diff_vs_same_bf16_weights": 0.0, | |
| "mean_abs_diff_vs_same_bf16_weights": 0.0, | |
| "max_abs_diff_vs_fp32_master": 0.18924713134765625, | |
| "mean_abs_diff_vs_fp32_master": 0.02142287604510784, | |
| "argmax_agree_vs_same_bf16_weights": 1.0, | |
| "argmax_agree_vs_fp32_master": 0.9983844757080078, | |
| "logit_abs_max": 30.520248413085938 | |
| }, | |
| { | |
| "input": "owm_val_clean", | |
| "tokens": 619, | |
| "max_abs_diff_vs_same_bf16_weights": 0.0, | |
| "mean_abs_diff_vs_same_bf16_weights": 0.0, | |
| "max_abs_diff_vs_fp32_master": 0.18924713134765625, | |
| "mean_abs_diff_vs_fp32_master": 0.02142287604510784, | |
| "argmax_agree_vs_same_bf16_weights": 1.0, | |
| "argmax_agree_vs_fp32_master": 0.9983844757080078, | |
| "logit_abs_max": 30.520248413085938 | |
| }, | |
| { | |
| "input": "gsm8k_test", | |
| "tokens": 121, | |
| "max_abs_diff_vs_same_bf16_weights": 0.0, | |
| "mean_abs_diff_vs_same_bf16_weights": 0.0, | |
| "max_abs_diff_vs_fp32_master": 0.17910480499267578, | |
| "mean_abs_diff_vs_fp32_master": 0.02021566778421402, | |
| "argmax_agree_vs_same_bf16_weights": 1.0, | |
| "argmax_agree_vs_fp32_master": 0.9834710955619812, | |
| "logit_abs_max": 22.752124786376953 | |
| }, | |
| { | |
| "input": "math_test", | |
| "tokens": 104, | |
| "max_abs_diff_vs_same_bf16_weights": 0.0, | |
| "mean_abs_diff_vs_same_bf16_weights": 0.0, | |
| "max_abs_diff_vs_fp32_master": 0.1441974639892578, | |
| "mean_abs_diff_vs_fp32_master": 0.021314028650522232, | |
| "argmax_agree_vs_same_bf16_weights": 1.0, | |
| "argmax_agree_vs_fp32_master": 1.0, | |
| "logit_abs_max": 25.434036254882812 | |
| }, | |
| { | |
| "input": "math_test_clean", | |
| "tokens": 104, | |
| "max_abs_diff_vs_same_bf16_weights": 0.0, | |
| "mean_abs_diff_vs_same_bf16_weights": 0.0, | |
| "max_abs_diff_vs_fp32_master": 0.1441974639892578, | |
| "mean_abs_diff_vs_fp32_master": 0.021314028650522232, | |
| "argmax_agree_vs_same_bf16_weights": 1.0, | |
| "argmax_agree_vs_fp32_master": 1.0, | |
| "logit_abs_max": 25.434036254882812 | |
| }, | |
| { | |
| "input": "random_1024", | |
| "tokens": 1024, | |
| "max_abs_diff_vs_same_bf16_weights": 0.0, | |
| "mean_abs_diff_vs_same_bf16_weights": 0.0, | |
| "max_abs_diff_vs_fp32_master": 0.2052001953125, | |
| "mean_abs_diff_vs_fp32_master": 0.024474455043673515, | |
| "argmax_agree_vs_same_bf16_weights": 1.0, | |
| "argmax_agree_vs_fp32_master": 0.994140625, | |
| "logit_abs_max": 17.417985916137695 | |
| } | |
| ], | |
| "logits_max_abs_diff_impl": 0.0, | |
| "logits_max_abs_diff_bf16_storage": 0.2052001953125, | |
| "left_padding_max_abs_diff": 3.147125244140625e-05, | |
| "tokenizer_ids_identical": "100/100", | |
| "tokenizer_roundtrip_exact": "100/100", | |
| "bpb_sample": { | |
| "records_per_set": 20, | |
| "max_chars_per_record": 20000, | |
| "block": 1024, | |
| "stride": 512, | |
| "per_set": { | |
| "owm_val": { | |
| "records": 20, | |
| "tokens": 19797, | |
| "utf8_bytes": 67807, | |
| "original_fp32_master": 1.04333, | |
| "hf_bf16_weights_fp32_compute": 1.04337, | |
| "rel_diff": 3.7595580306251906e-05 | |
| }, | |
| "owm_val_clean": { | |
| "records": 20, | |
| "tokens": 21497, | |
| "utf8_bytes": 80082, | |
| "original_fp32_master": 1.13213, | |
| "hf_bf16_weights_fp32_compute": 1.13217, | |
| "rel_diff": 3.843563384466128e-05 | |
| }, | |
| "gsm8k_test": { | |
| "records": 20, | |
| "tokens": 3636, | |
| "utf8_bytes": 10910, | |
| "original_fp32_master": 1.04602, | |
| "hf_bf16_weights_fp32_compute": 1.04612, | |
| "rel_diff": 9.461596812371703e-05 | |
| }, | |
| "math_test": { | |
| "records": 20, | |
| "tokens": 4062, | |
| "utf8_bytes": 10074, | |
| "original_fp32_master": 0.8763, | |
| "hf_bf16_weights_fp32_compute": 0.87655, | |
| "rel_diff": 0.00027754500238853655 | |
| }, | |
| "math_test_clean": { | |
| "records": 20, | |
| "tokens": 3983, | |
| "utf8_bytes": 9958, | |
| "original_fp32_master": 0.89883, | |
| "hf_bf16_weights_fp32_compute": 0.89907, | |
| "rel_diff": 0.0002619447947951707 | |
| } | |
| }, | |
| "pooled_original_fp32_master": 1.0658, | |
| "pooled_hf_bf16_weights_fp32_compute": 1.06587, | |
| "pooled_rel_diff": 6.305839549519214e-05, | |
| "seconds": 107.1 | |
| }, | |
| "full_set": { | |
| "set": "gsm8k_test", | |
| "records": 1319, | |
| "tokens": 225351, | |
| "utf8_bytes": 652230, | |
| "frozen_bytes": 652230, | |
| "hf_bpb": 1.07, | |
| "frozen_bpb_gpu_bf16_autocast": 1.07016, | |
| "rel_diff": -0.00014918026252958935, | |
| "note": "", | |
| "seconds": 217.3 | |
| } | |
| } |