Instructions to use ecloudtech/Erk-14B with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use ecloudtech/Erk-14B with Transformers:

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

pipe = pipeline("text-generation", model="ecloudtech/Erk-14B")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# pip install -U transformers accelerate
# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("ecloudtech/Erk-14B")
model = AutoModelForCausalLM.from_pretrained("ecloudtech/Erk-14B", device_map="auto")
messages = [
    {"role": "user", "content": "Who are you?"},
]
inputs = tokenizer.apply_chat_template(
	messages,
	add_generation_prompt=True,
	tokenize=True,
	return_dict=True,
	return_tensors="pt",
).to(model.device)

outputs = model.generate(**inputs, max_new_tokens=256)
print(tokenizer.decode(outputs[0][inputs["input_ids"].shape[-1]:]))

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use ecloudtech/Erk-14B with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "ecloudtech/Erk-14B"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "ecloudtech/Erk-14B",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/ecloudtech/Erk-14B

SGLang

How to use ecloudtech/Erk-14B 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 "ecloudtech/Erk-14B" \
    --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": "ecloudtech/Erk-14B",
		"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 "ecloudtech/Erk-14B" \
        --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": "ecloudtech/Erk-14B",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use ecloudtech/Erk-14B with Docker Model Runner:
```
docker model run hf.co/ecloudtech/Erk-14B
```

aigencydev commited on Sep 8

Commit

3be9b60

verified ·

1 Parent(s): 2ab0b27

TurkishMMLU: kirlilik disi, sik dondurmeli protokolde ek olcum (%67,18); ayrinti Erk-32B raporunda

Browse files

Files changed (1) hide show

README.md +40 -0

README.md CHANGED Viewed

@@ -213,6 +213,46 @@ büyük kapasite, daha derin muhakeme ve alan uzmanlığı hedefleniyor.
 ---
 ## English
 ### What is Erk?

 ---
+## Ek ölçüm: kirlilik dışı, şık döndürmeli protokol (Eylül 2026)
+Yukarıdaki %69,7, TurkishMMLU'nun tamamında, lm-evaluation-harness ile 0-shot,
+şık sırası döndürülmeden ölçülmüştür. Erk-32B çalışmasında geliştirdiğimiz daha
+sıkı protokolle aynı model yeniden ölçüldü:
+- Test kümesinin 900 sorusundan **113'ü (%12,6)** devam-eğitimi külliyatında
+  birebir geçiyor ([liste](https://huggingface.co/datasets/ecloudtech/TurkishMMLU-Kirlilik));
+  bunlar ve eğitim sırasında izlenen 150 soru dışarıda bırakıldı → **temiz 652**.
+- Her soru şık sırası döndürülerek ölçüldü (harf/konum yanlılığı elenir).
+- Fark, 10.000 yeniden örneklemeli eşli bootstrap %95 güven aralığıyla verildi.
+| Model | Temiz 652, döndürülmüş |
+|---|---|
+| Qwen3-14B (taban) | %60,58 |
+| **Erk-14B** | **%67,18** |
+| Fark | **+6,60 [+3,07, +10,28]** |
+İki sayı çelişmez: %69,7 ile %67,18 aynı modelin iki farklı koşulda ölçümüdür;
+fark protokolden gelir. Tabana göre kazanç her iki koşulda da anlamlıdır
+(+6,3 tam küme, +6,6 temiz küme). Yukarıdaki sıralama tablosu tam-küme
+protokolüne aittir; karşılaştırılan diğer modeller temiz protokolle yeniden
+ölçülmemiştir.
+Protokolün tamamı ve yeniden üretim betikleri:
+[Erk-32B — Ölçüm Protokolü](https://huggingface.co/ecloudtech/Erk-32B).
+### English
+The 69.7% above was measured on the full TurkishMMLU test set with
+lm-evaluation-harness, 0-shot, without option rotation. Under the stricter
+protocol developed for Erk-32B — 113/900 questions found verbatim in the
+pretraining corpus excluded (plus 150 used for monitoring), remaining 652
+measured with option rotation, paired bootstrap CIs — Erk-14B scores
+**67.18%** vs **60.58%** for Qwen3-14B (+6.60 [+3.07, +10.28]). The two numbers
+are the same model under two protocols; the gain over the base is significant
+under both. The ranking table above is on the full-set protocol; other models
+were not re-measured under the clean one. Full protocol and scripts:
+[Erk-32B](https://huggingface.co/ecloudtech/Erk-32B).
 ## English
 ### What is Erk?