Text Generation
Transformers
PyTorch
TensorBoard
gpt2
Generated from Trainer
custom_code
text-generation-inference
Instructions to use mrm8488/santacoder-finetuned-the-stack-dockerfiles with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Transformers
How to use mrm8488/santacoder-finetuned-the-stack-dockerfiles with Transformers:
# Use a pipeline as a high-level helper from transformers import pipeline pipe = pipeline("text-generation", model="mrm8488/santacoder-finetuned-the-stack-dockerfiles", trust_remote_code=True)# Load model directly from transformers import AutoTokenizer, AutoModelForCausalLM tokenizer = AutoTokenizer.from_pretrained("mrm8488/santacoder-finetuned-the-stack-dockerfiles", trust_remote_code=True) model = AutoModelForCausalLM.from_pretrained("mrm8488/santacoder-finetuned-the-stack-dockerfiles", trust_remote_code=True) - Notebooks
- Google Colab
- Kaggle
- Local Apps Settings
- vLLM
How to use mrm8488/santacoder-finetuned-the-stack-dockerfiles with vLLM:
Install from pip and serve model
# Install vLLM from pip: pip install vllm # Start the vLLM server: vllm serve "mrm8488/santacoder-finetuned-the-stack-dockerfiles" # Call the server using curl (OpenAI-compatible API): curl -X POST "http://localhost:8000/v1/completions" \ -H "Content-Type: application/json" \ --data '{ "model": "mrm8488/santacoder-finetuned-the-stack-dockerfiles", "prompt": "Once upon a time,", "max_tokens": 512, "temperature": 0.5 }'Use Docker
docker model run hf.co/mrm8488/santacoder-finetuned-the-stack-dockerfiles
- SGLang
How to use mrm8488/santacoder-finetuned-the-stack-dockerfiles 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 "mrm8488/santacoder-finetuned-the-stack-dockerfiles" \ --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": "mrm8488/santacoder-finetuned-the-stack-dockerfiles", "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 "mrm8488/santacoder-finetuned-the-stack-dockerfiles" \ --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": "mrm8488/santacoder-finetuned-the-stack-dockerfiles", "prompt": "Once upon a time,", "max_tokens": 512, "temperature": 0.5 }' - Docker Model Runner
How to use mrm8488/santacoder-finetuned-the-stack-dockerfiles with Docker Model Runner:
docker model run hf.co/mrm8488/santacoder-finetuned-the-stack-dockerfiles
| FROM python:3.12-slim | |
| WORKDIR /app | |
| COPY requirements.txt . | |
| COPY app.py . | |
| RUN pip install --upgrade pip | |
| RUN pip install --no-cache-dir -r requirements.txt | |
| RUN pip install gunicorn | |
| EXPOSE 7860 | |
| CMD ["gunicorn", "app:app", \ | |
| "-w", "4", \ | |
| "--worker-class", "gevent", \ | |
| "--worker-connections", "100", \ | |
| "-b", "0.0.0.0:7860", \ | |
| "--timeout", "120", \ | |
| "--keep-alive", "5", \ | |
| "--max-requests", "1000", \ | |
| "--max-requests-jitter", "100"] | |
| app.py dosyası oluşturun | |
| from flask import Flask, request, Response | |
| import requests | |
| from urllib.parse import urlparse, urljoin, quote, unquote | |
| import re | |
| import traceback | |
| import os | |
| app = Flask(__name__) | |
| def detect_m3u_type(content): | |
| """Rileva se è un M3U (lista IPTV) o un M3U8 (flusso HLS)""" | |
| if "#EXTM3U" in content and "#EXTINF" in content: | |
| return "m3u8" | |
| return "m3u" | |
| def replace_key_uri(line, headers_query): | |
| """Sostituisce l'URI della chiave AES-128 con il proxy""" | |
| match = re.search(r'URI="([^"]+)"', line) | |
| if match: | |
| key_url = match.group(1) | |
| proxied_key_url = f"/proxy/key?url={quote(key_url)}&{headers_query}" | |
| return line.replace(key_url, proxied_key_url) | |
| return line | |
| def resolve_m3u8_link(url, headers=None): | |
| """ | |
| Tenta di risolvere un URL M3U8. | |
| Prova prima la logica specifica per iframe (tipo Daddylive), inclusa la lookup della server_key. | |
| Se fallisce, verifica se l'URL iniziale era un M3U8 diretto e lo restituisce. | |
| """ | |
| if not url: | |
| print("Errore: URL non fornito.") | |
| return {"resolved_url": None, "headers": {}} | |
| print(f"Tentativo di risoluzione URL: {url}") | |
| # Utilizza gli header forniti, altrimenti usa un User-Agent di default | |
| current_headers = headers if headers else {'User-Agent': 'Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/133.0.0.0+Safari/537.36'} | |
| initial_response_text = None | |
| final_url_after_redirects = None | |
| # Verifica se è un URL di vavoo.to | |
| is_vavoo = "vavoo.to" in url.lower() | |
| try: | |
| with requests.Session() as session: | |
| print(f"Passo 1: Richiesta a {url}") | |
| response = session.get(url, headers=current_headers, allow_redirects=True, timeout=(5, 15)) | |
| response.raise_for_status() | |
| initial_response_text = response.text | |
| final_url_after_redirects = response.url | |
| print(f"Passo 1 completato. URL finale dopo redirect: {final_url_after_redirects}") | |
| # Se è un URL di vavoo.to, salta la logica dell'iframe | |
| if is_vavoo: | |
| if initial_response_text and initial_response_text.strip().startswith('#EXTM3U'): | |
| return { | |
| "resolved_url": final_url_after_redirects, | |
| "headers": current_headers | |
| } | |
| else: | |
| # Se non è un M3U8 diretto, restituisci l'URL originale per vavoo | |
| print(f"URL vavoo.to non è un M3U8 diretto: {url}") | |
| return { | |
| "resolved_url": url, | |
| "headers": current_headers | |
| } | |
| # Prova la logica dell'iframe per gli altri URL | |
| print("Tentativo con logica iframe...") | |
| try: | |
| # Secondo passo (Iframe): Trova l'iframe src nella risposta iniziale | |
| iframes = re.findall(r'iframe src="([^"]+)"', initial_response_text) | |
| if not iframes: | |
| raise ValueError("Nessun iframe src trovato.") | |
| url2 = iframes[0] | |
| print(f"Passo 2 (Iframe): Trovato iframe URL: {url2}") | |
| # Terzo passo (Iframe): Richiesta all'URL dell'iframe | |
| referer_raw = urlparse(url2).scheme + "://" + urlparse(url2).netloc + "/" | |
| origin_raw = urlparse(url2).scheme + "://" + urlparse(url2).netloc | |
| current_headers['Referer'] = referer_raw | |
| current_headers['Origin'] = origin_raw | |
| print(f"Passo 3 (Iframe): Richiesta a {url2}") | |
| response = session.get(url2, headers=current_headers, timeout=(5, 15)) | |
| response.raise_for_status() | |
| iframe_response_text = response.text | |
| print("Passo 3 (Iframe) completato.") | |
| # Quarto passo (Iframe): Estrai parametri dinamici dall'iframe response | |
| channel_key_match = re.search(r'(?s) channelKey = \"([^\"]*)"', iframe_response_text) | |
| auth_ts_match = re.search(r'(?s) authTs\s*= \"([^\"]*)"', iframe_response_text) | |
| auth_rnd_match = re.search(r'(?s) authRnd\s*= \"([^\"]*)"', iframe_response_text) | |
| auth_sig_match = re.search(r'(?s) authSig\s*= \"([^\"]*)"', iframe_response_text) | |
| auth_host_match = re.search(r'\}\s*fetchWithRetry\(\s*\'([^\']*)\'', iframe_response_text) | |
| server_lookup_match = re.search(r'n fetchWithRetry\(\s*\'([^\']*)\'', iframe_response_text) | |
| if not all([channel_key_match, auth_ts_match, auth_rnd_match, auth_sig_match, auth_host_match, server_lookup_match]): | |
| raise ValueError("Impossibile estrarre tutti i parametri dinamici dall'iframe response.") | |
| channel_key = channel_key_match.group(1) | |
| auth_ts = auth_ts_match.group(1) | |
| auth_rnd = auth_rnd_match.group(1) | |
| auth_sig = quote(auth_sig_match.group(1)) | |
| auth_host = auth_host_match.group(1) | |
| server_lookup = server_lookup_match.group(1) | |
| print("Passo 4 (Iframe): Parametri dinamici estratti.") | |
| # Quinto passo (Iframe): Richiesta di autenticazione | |
| auth_url = f'{auth_host}{channel_key}&ts={auth_ts}&rnd={auth_rnd}&sig={auth_sig}' | |
| print(f"Passo 5 (Iframe): Richiesta di autenticazione a {auth_url}") | |
| auth_response = session.get(auth_url, headers=current_headers, timeout=(5, 15)) | |
| auth_response.raise_for_status() | |
| print("Passo 5 (Iframe) completato.") | |
| # Sesto passo (Iframe): Richiesta di server lookup per ottenere la server_key | |
| server_lookup_url = f"https://{urlparse(url2).netloc}{server_lookup}{channel_key}" | |
| print(f"Passo 6 (Iframe): Richiesta server lookup a {server_lookup_url}") | |
| server_lookup_response = session.get(server_lookup_url, headers=current_headers, timeout=(5, 15)) | |
| server_lookup_response.raise_for_status() | |
| server_lookup_data = server_lookup_response.json() | |
| print("Passo 6 (Iframe) completato.") | |
| # Settimo passo (Iframe): Estrai server_key dalla risposta di server lookup | |
| server_key = server_lookup_data.get('server_key') | |
| if not server_key: | |
| raise ValueError("'server_key' non trovato nella risposta di server lookup.") | |
| print(f"Passo 7 (Iframe): Estratto server_key: {server_key}") | |
| # Ottavo passo (Iframe): Costruisci il link finale | |
| host_match = re.search('(?s)m3u8 =.*?:.*?:.*?".*?".*?"([^"]*)"', iframe_response_text) | |
| if not host_match: | |
| raise ValueError("Impossibile trovare l'host finale per l'm3u8.") | |
| host = host_match.group(1) | |
| print(f"Passo 8 (Iframe): Trovato host finale per m3u8: {host}") | |
| # Costruisci l'URL finale del flusso | |
| final_stream_url = ( | |
| f'https://{server_key}{host}{server_key}/{channel_key}/mono.m3u8' | |
| ) | |
| # Prepara gli header per lo streaming | |
| stream_headers = { | |
| 'User-Agent': current_headers.get('User-Agent', ''), | |
| 'Referer': referer_raw, | |
| 'Origin': origin_raw | |
| } | |
| return { | |
| "resolved_url": final_stream_url, | |
| "headers": stream_headers | |
| } | |
| except (ValueError, requests.exceptions.RequestException) as e: | |
| print(f"Logica iframe fallita: {e}") | |
| print("Tentativo fallback: verifica se l'URL iniziale era un M3U8 diretto...") | |
| # Fallback: Verifica se la risposta iniziale era un file M3U8 diretto | |
| if initial_response_text and initial_response_text.strip().startswith('#EXTM3U'): | |
| print("Fallback riuscito: Trovato file M3U8 diretto.") | |
| return { | |
| "resolved_url": final_url_after_redirects, | |
| "headers": current_headers | |
| } | |
| else: | |
| print("Fallback fallito: La risposta iniziale non era un M3U8 diretto.") | |
| return { | |
| "resolved_url": url, | |
| "headers": current_headers | |
| } | |
| except requests.exceptions.RequestException as e: | |
| print(f"Errore durante la richiesta HTTP iniziale: {e}") | |
| return {"resolved_url": url, "headers": current_headers} | |
| except Exception as e: | |
| print(f"Errore generico durante la risoluzione: {e}") | |
| return {"resolved_url": url, "headers": current_headers} | |
| def proxy(): | |
| """Proxy per liste M3U che aggiunge automaticamente /proxy/m3u?url= con IP prima dei link""" | |
| m3u_url = request.args.get('url', '').strip() | |
| if not m3u_url: | |
| return "Errore: Parametro 'url' mancante", 400 | |
| try: | |
| # Ottieni l'IP del server | |
| server_ip = request.host | |
| # Scarica la lista M3U originale | |
| response = requests.get(m3u_url, timeout=(10, 30)) # Timeout connessione 10s, lettura 30s | |
| response.raise_for_status() | |
| m3u_content = response.text | |
| # Modifica solo le righe che contengono URL (non iniziano con #) | |
| modified_lines = [] | |
| for line in m3u_content.splitlines(): | |
| line = line.strip() | |
| if line and not line.startswith('#'): | |
| # Per tutti i link, usa il proxy normale | |
| modified_line = f"http://{server_ip}/proxy/m3u?url={line}" | |
| modified_lines.append(modified_line) | |
| else: | |
| # Mantieni invariate le righe di metadati | |
| modified_lines.append(line) | |
| modified_content = '\n'.join(modified_lines) | |
| # Estrai il nome del file dall'URL originale | |
| parsed_m3u_url = urlparse(m3u_url) | |
| original_filename = os.path.basename(parsed_m3u_url.path) | |
| return Response(modified_content, content_type="application/vnd.apple.mpegurl", headers={'Content-Disposition': f'attachment; filename="{original_filename}"'}) | |
| except requests.RequestException as e: | |
| return f"Errore durante il download della lista M3U: {str(e)}", 500 | |
| except Exception as e: | |
| return f"Errore generico: {str(e)}", 500 | |
| def proxy_m3u(): | |
| """Proxy per file M3U e M3U8 con supporto per redirezioni e header personalizzati""" | |
| m3u_url = request.args.get('url', '').strip() | |
| if not m3u_url: | |
| return "Errore: Parametro 'url' mancante", 400 | |
| default_headers = { | |
| "User-Agent": "Mozilla/5.0 (iPhone; CPU iPhone OS 14_5 like Mac OS X) AppleWebKit/605.1.15 (KHTML, like Gecko) FxiOS/33.0 Mobile/15E148 Safari/605.1.15", | |
| "Referer": "https://vavoo.to/", | |
| "Origin": "https://vavoo.to" | |
| } | |
| # Estrai gli header dalla richiesta, sovrascrivendo i default | |
| request_headers = { | |
| unquote(key[2:]).replace("_", "-"): unquote(value).strip() | |
| for key, value in request.args.items() | |
| if key.lower().startswith("h_") | |
| } | |
| headers = {**default_headers, **request_headers} | |
| # --- Logica per trasformare l'URL se necessario --- | |
| processed_url = m3u_url | |
| # Trasforma /stream/ in /embed/ per Daddylive | |
| if '/stream/stream-' in m3u_url and 'daddylive.dad' in m3u_url: | |
| processed_url = m3u_url.replace('/stream/stream-', '/embed/stream-') | |
| print(f"URL {m3u_url} trasformato da /stream/ a /embed/: {processed_url}") | |
| match_premium_m3u8 = re.search(r'/premium(\d+)/mono\.m3u8$', m3u_url) | |
| if match_premium_m3u8: | |
| channel_number = match_premium_m3u8.group(1) | |
| transformed_url = f"https://daddylive.dad/embed/stream-{channel_number}.php" | |
| print(f"URL {m3u_url} corrisponde al pattern premium. Trasformato in: {transformed_url}") | |
| processed_url = transformed_url | |
| else: | |
| print(f"URL {processed_url} processato per la risoluzione.") | |
| try: | |
| print(f"Chiamata a resolve_m3u8_link per URL processato: {processed_url}") | |
| result = resolve_m3u8_link(processed_url, headers) | |
| if not result["resolved_url"]: | |
| return "Errore: Impossibile risolvere l'URL in un M3U8 valido.", 500 | |
| resolved_url = result["resolved_url"] | |
| current_headers_for_proxy = result["headers"] | |
| print(f"Risoluzione completata. URL M3U8 finale: {resolved_url}") | |
| # Fetchare il contenuto M3U8 effettivo dall'URL risolto | |
| print(f"Fetching M3U8 content from resolved URL: {resolved_url}") | |
| m3u_response = requests.get(resolved_url, headers=current_headers_for_proxy, allow_redirects=True, timeout=(10, 20)) # Timeout connessione 10s, lettura 20s | |
| m3u_response.raise_for_status() | |
| m3u_content = m3u_response.text | |
| final_url = m3u_response.url | |
| # Processa il contenuto M3U8 | |
| file_type = detect_m3u_type(m3u_content) | |
| if file_type == "m3u": | |
| return Response(m3u_content, content_type="application/vnd.apple.mpegurl") | |
| # Processa contenuto M3U8 | |
| parsed_url = urlparse(final_url) | |
| base_url = f"{parsed_url.scheme}://{parsed_url.netloc}{parsed_url.path.rsplit('/', 1)[0]}/" | |
| # Prepara la query degli header per segmenti/chiavi proxati | |
| headers_query = "&".join([f"h_{quote(k)}={quote(v)}" for k, v in current_headers_for_proxy.items()]) | |
| modified_m3u8 = [] | |
| for line in m3u_content.splitlines(): | |
| line = line.strip() | |
| if line.startswith("#EXT-X-KEY") and 'URI="' in line: | |
| line = replace_key_uri(line, headers_query) | |
| elif line and not line.startswith("#"): | |
| segment_url = urljoin(base_url, line) | |
| line = f"/proxy/ts?url={quote(segment_url)}&{headers_query}" | |
| modified_m3u8.append(line) | |
| modified_m3u8_content = "\n".join(modified_m3u8) | |
| return Response(modified_m3u8_content, content_type="application/vnd.apple.mpegurl") | |
| except requests.RequestException as e: | |
| print(f"Errore durante il download o la risoluzione del file: {str(e)}") | |
| return f"Errore durante il download o la risoluzione del file M3U/M3U8: {str(e)}", 500 | |
| except Exception as e: | |
| print(f"Errore generico nella funzione proxy_m3u: {str(e)}") | |
| return f"Errore generico durante l'elaborazione: {str(e)}", 500 | |
| def proxy_resolve(): | |
| """Proxy per risolvere e restituire un URL M3U8""" | |
| url = request.args.get('url', '').strip() | |
| if not url: | |
| return "Errore: Parametro 'url' mancante", 400 | |
| headers = { | |
| unquote(key[2:]).replace("_", "-"): unquote(value).strip() | |
| for key, value in request.args.items() | |
| if key.lower().startswith("h_") | |
| } | |
| try: | |
| result = resolve_m3u8_link(url, headers) | |
| if not result["resolved_url"]: | |
| return "Errore: Impossibile risolvere l'URL", 500 | |
| headers_query = "&".join([f"h_{quote(k)}={quote(v)}" for k, v in result["headers"].items()]) | |
| return Response( | |
| f"#EXTM3U\n" | |
| f"#EXTINF:-1,Canale Risolto\n" | |
| f"/proxy/m3u?url={quote(result['resolved_url'])}&{headers_query}", | |
| content_type="application/vnd.apple.mpegurl" | |
| ) | |
| except Exception as e: | |
| return f"Errore durante la risoluzione dell'URL: {str(e)}", 500 | |
| def proxy_ts(): | |
| """Proxy per segmenti .TS con headers personalizzati - SENZA CACHE""" | |
| ts_url = request.args.get('url', '').strip() | |
| if not ts_url: | |
| return "Errore: Parametro 'url' mancante", 400 | |
| headers = { | |
| unquote(key[2:]).replace("_", "-"): unquote(value).strip() | |
| for key, value in request.args.items() | |
| if key.lower().startswith("h_") | |
| } | |
| try: | |
| # Stream diretto senza cache per evitare freezing | |
| response = requests.get(ts_url, headers=headers, stream=True, allow_redirects=True, timeout=(10, 30)) # Timeout di connessione 10s, lettura 30s | |
| response.raise_for_status() | |
| def generate(): | |
| for chunk in response.iter_content(chunk_size=8192): | |
| if chunk: | |
| yield chunk | |
| return Response(generate(), content_type="video/mp2t") | |
| except requests.RequestException as e: | |
| return f"Errore durante il download del segmento TS: {str(e)}", 500 | |
| def proxy_key(): | |
| """Proxy per la chiave AES-128 con header personalizzati""" | |
| key_url = request.args.get('url', '').strip() | |
| if not key_url: | |
| return "Errore: Parametro 'url' mancante per la chiave", 400 | |
| headers = { | |
| unquote(key[2:]).replace("_", "-"): unquote(value).strip() | |
| for key, value in request.args.items() | |
| if key.lower().startswith("h_") | |
| } | |
| try: | |
| response = requests.get(key_url, headers=headers, allow_redirects=True, timeout=(5, 15)) # Timeout connessione 5s, lettura 15s | |
| response.raise_for_status() | |
| return Response(response.content, content_type="application/octet-stream") | |
| except requests.RequestException as e: | |
| return f"Errore durante il download della chiave AES-128: {str(e)}", 500 | |
| def index(): | |
| """Pagina principale che mostra un messaggio di benvenuto""" | |
| return "Proxy started!" | |
| if __name__ == '__main__': | |
| print("Proxy started!") | |
| app.run(host="0.0.0.0", port=7860, debug=False) |