Instructions to use Maincode/Matilda-Platypus-13B-Base with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use Maincode/Matilda-Platypus-13B-Base with Transformers:

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

pipe = pipeline("text-generation", model="Maincode/Matilda-Platypus-13B-Base", trust_remote_code=True)

# pip install -U transformers accelerate
# Load model directly
from transformers import AutoModelForCausalLM
model = AutoModelForCausalLM.from_pretrained("Maincode/Matilda-Platypus-13B-Base", trust_remote_code=True, device_map="auto")

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use Maincode/Matilda-Platypus-13B-Base with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "Maincode/Matilda-Platypus-13B-Base"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "Maincode/Matilda-Platypus-13B-Base",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Use Docker

docker model run hf.co/Maincode/Matilda-Platypus-13B-Base

SGLang

How to use Maincode/Matilda-Platypus-13B-Base 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 "Maincode/Matilda-Platypus-13B-Base" \
    --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": "Maincode/Matilda-Platypus-13B-Base",
		"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 "Maincode/Matilda-Platypus-13B-Base" \
        --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": "Maincode/Matilda-Platypus-13B-Base",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Docker Model Runner
How to use Maincode/Matilda-Platypus-13B-Base with Docker Model Runner:
```
docker model run hf.co/Maincode/Matilda-Platypus-13B-Base
```

yue-maincode commited on 4 days ago

Commit

a168a7c

verified ·

1 Parent(s): 796c698

Update README.md (whole-number model naming)

Browse files

Files changed (1) hide show

README.md +4 -4

README.md CHANGED Viewed

@@ -12,11 +12,11 @@ tags:
 - qwen3
 - base-model
 ---
-# Matilda Platypus 13.5B (base)
 **Any size you want.** Matilda Platypus is a family of intermediate-size language models built by connecting the lower
 layers of a larger pretrained Qwen3 model to the upper layers of a smaller one. A learned linear **Representation
-Bridge** aligns the hidden representations at the connection. This checkpoint is the **13.5B** member
 (cut `29+13`): 13.51B parameters, 26.3 GFLOP per token, 91% of the
 larger parent's compute. It is a base model (no instruction tuning); see the technical report (`technical_report.pdf`)
 for the method and the full evaluation.
@@ -39,7 +39,7 @@ Base-model evaluation with lm-eval-harness 0.4.13, identical protocol for the mo
 multiple choice, MMLU 5-shot, GSM8K 8-shot chain-of-thought, BBH 3-shot chain-of-thought, HumanEval 0-shot and MBPP
 3-shot with the generated code executed; HumanEval scored with markdown fences stripped). Scores are proportions.
-| Benchmark | Qwen3-8B-Base | **Platypus 13.5B** | Qwen3-14B-Base |
 |---|---|---|---|
 | HumanEval pass@1 | 0.677 | **0.750** | 0.720 |
 | GSM8K (8-shot CoT) | 0.856 | **0.882** | 0.870 |
@@ -75,7 +75,7 @@ parent; it is not a verified latency or cost measurement. Full tables and figure
 import torch
 from transformers import AutoModelForCausalLM, AutoTokenizer
-repo = "Maincode/matilda-platypus-13.5b-base"
 tok = AutoTokenizer.from_pretrained(repo)
 model = AutoModelForCausalLM.from_pretrained(repo, trust_remote_code=True, dtype=torch.bfloat16).cuda()
 inputs = tok("The capital of Australia is", return_tensors="pt").to("cuda")

 - qwen3
 - base-model
 ---
+# Matilda Platypus 13B (base)
 **Any size you want.** Matilda Platypus is a family of intermediate-size language models built by connecting the lower
 layers of a larger pretrained Qwen3 model to the upper layers of a smaller one. A learned linear **Representation
+Bridge** aligns the hidden representations at the connection. This checkpoint is the **13B** member
 (cut `29+13`): 13.51B parameters, 26.3 GFLOP per token, 91% of the
 larger parent's compute. It is a base model (no instruction tuning); see the technical report (`technical_report.pdf`)
 for the method and the full evaluation.
 multiple choice, MMLU 5-shot, GSM8K 8-shot chain-of-thought, BBH 3-shot chain-of-thought, HumanEval 0-shot and MBPP
 3-shot with the generated code executed; HumanEval scored with markdown fences stripped). Scores are proportions.
+| Benchmark | Qwen3-8B-Base | **Platypus 13B** | Qwen3-14B-Base |
 |---|---|---|---|
 | HumanEval pass@1 | 0.677 | **0.750** | 0.720 |
 | GSM8K (8-shot CoT) | 0.856 | **0.882** | 0.870 |
 import torch
 from transformers import AutoModelForCausalLM, AutoTokenizer
+repo = "Maincode/Matilda-Platypus-13B-Base"
 tok = AutoTokenizer.from_pretrained(repo)
 model = AutoModelForCausalLM.from_pretrained(repo, trust_remote_code=True, dtype=torch.bfloat16).cuda()
 inputs = tok("The capital of Australia is", return_tensors="pt").to("cuda")