Text Generation
Transformers
Safetensors
Turkish
qwen3
turkish
türkçe
continued-pretraining
turkish-llm
türkçe-dil-modeli
conversational
Eval Results (legacy)
text-generation-inference
Instructions to use ecloudtech/Erk-32B with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Transformers
How to use ecloudtech/Erk-32B with Transformers:
# Use a pipeline as a high-level helper from transformers import pipeline pipe = pipeline("text-generation", model="ecloudtech/Erk-32B") 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-32B") model = AutoModelForCausalLM.from_pretrained("ecloudtech/Erk-32B", 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-32B with vLLM:
Install from pip and serve model
# Install vLLM from pip: pip install vllm # Start the vLLM server: vllm serve "ecloudtech/Erk-32B" # 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-32B", "messages": [ { "role": "user", "content": "What is the capital of France?" } ] }'Use Docker
docker model run hf.co/ecloudtech/Erk-32B
- SGLang
How to use ecloudtech/Erk-32B 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-32B" \ --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-32B", "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-32B" \ --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-32B", "messages": [ { "role": "user", "content": "What is the capital of France?" } ] }' - Docker Model Runner
How to use ecloudtech/Erk-32B with Docker Model Runner:
docker model run hf.co/ecloudtech/Erk-32B
Download betikler/bootstrap_sft2.py from ecloudtech/Erk-32B: direct link, hf CLI and curl.
- Browser
- Download file 1.95 kB
-
https://huggingface.co/ecloudtech/Erk-32B/resolve/main/betikler/bootstrap_sft2.py
- Command line
-
hf download hf://ecloudtech/Erk-32B/betikler/bootstrap_sft2.py
-
curl -L -o bootstrap_sft2.py https://huggingface.co/ecloudtech/Erk-32B/resolve/main/betikler/bootstrap_sft2.py
1.95 kB
| """Ek tur sonrasi olcek taramasi — esli bootstrap guven araliklari. | |
| Soru-bazli vektorler ayni sorular uzerinde oldugu icin fark ESLI hesaplanir: | |
| her yeniden orneklemede ayni soru indeksleri iki kosula da uygulanir. Bagimsiz | |
| bootstrap'a gore araliklar daha dar ve dogru olur (soru zorlugu ortak varyans). | |
| 10.000 yeniden ornekleme, tohum 0 — OLCUM-PROTOKOLU ile ayni. | |
| """ | |
| import json, numpy as np | |
| A = json.load(open("/ari/users/ytanis/olcek_ince_sft2.json")) # ek tur SONRASI | |
| try: | |
| B = json.load(open("/ari/users/ytanis/olcek_ince.json")) # ek tur ONCESI | |
| except Exception: | |
| B = {} | |
| rng = np.random.default_rng(0) | |
| N = 10000 | |
| def ci(x, y): | |
| x = np.asarray(x, float); y = np.asarray(y, float); n = len(x) | |
| assert n == len(y) | |
| idx = rng.integers(0, n, size=(N, n)) | |
| d = (x[idx].mean(1) - y[idx].mean(1)) * 100 | |
| lo, hi = np.percentile(d, [2.5, 97.5]) | |
| return (x.mean() - y.mean()) * 100, lo, hi | |
| def satir(ad, x, y): | |
| f, lo, hi = ci(x, y) | |
| anl = "ANLAMLI" if (lo > 0 or hi < 0) else " - " | |
| print(" %-40s %+6.2f [%+6.2f, %+6.2f] %s" % (ad, f, lo, hi, anl)) | |
| print("EK TUR SONRASI — olcekler arasi (esli, n_sohbet=329, n_kiyas=652)") | |
| print(" SOHBET") | |
| for a, b in [("0.80", "0.00"), ("1.00", "0.00"), ("0.40", "0.00"), ("0.80", "0.40"), ("0.80", "1.00")]: | |
| satir("sohbet %s vs %s" % (a, b), A[a]["sohbet_vec"], A[b]["sohbet_vec"]) | |
| print(" KIYAS") | |
| for a, b in [("0.40", "0.00"), ("0.80", "0.00"), ("1.00", "0.00"), ("0.40", "0.80"), ("0.40", "1.00")]: | |
| satir("kiyas %s vs %s" % (a, b), A[a]["mmlu_vec"], A[b]["mmlu_vec"]) | |
| if B: | |
| print("\nEK TUR SONRASI vs ONCESI — ayni olcekte (esli)") | |
| for o in ["0.40", "0.50", "1.00"]: | |
| if o in B: | |
| satir("sohbet %s: sonra vs once" % o, A[o]["sohbet_vec"], B[o]["sohbet_vec"]) | |
| satir("kiyas %s: sonra vs once" % o, A[o]["mmlu_vec"], B[o]["mmlu_vec"]) | |
| else: | |
| print(" (%s once dosyasinda yok)" % o) | |