zoltanctoth

Solution	Transcription Speed (1hr audio)	Cost/hr	Hungarian Quality	Speaker Diarization	Notes
MacBook Air (local)	~45-60 min	Free	Good	Needs separate model (pyannote)	Slow, pyannote adds ~20-30 min extra per hour
OpenAI Whisper API	~1-3 min	$0.36	Good	❌ Not supported	Need to combine with separate diarization
Whisper + pyannote	~5-10 min (GPU)	Free	Good	✅ Yes (local)	Best free option, needs GPU + HuggingFace token
Deepgram	~30-60 sec	$0.25	Decent	✅ Built-in	Fast, easy API, HU quality not the best

zoltanctoth

Miért vagyunk ennyire ingerültek, amikor még soha nem volt ilyen jó dolgunk?

Dr. Ludwig Hasler előadása
Hittisauer Landgespräche - "Jólét, hogyan tovább"

---

Bevezetés

Lényegében Ausztriában éppen egy korábban sosem ismert anyagi jólétnek örvendünk. Ráadásul elérhető egészségügyi ellátásunk van, ingyenes iskolai oktatás, és biztonságban élünk.

zoltanctoth

Warum sind wir so gereizt, wo es uns doch besser geht als je?

Vortrag von Dr. Ludwig Hasler
Hittisauer Landgespräche - "Wohlstand, wie weiter"

---

Einleitung

Willkommen bei Fokus. Im Grunde erfreuen wir uns in Österreich doch gerade eines nie gekannten materiellen Wohlstands. Zudem haben wir eine leistbare Gesundheitsversorgung, eine kostenlose Schulbildung und wir leben in Sicherheit.

zoltanctoth

professionalism = mlflow.metrics.genai.make_genai_metric(
    name="professionalism",
    definition=(
        "Professionalism refers to the use of a formal, respectful, and appropriate style of communication that is "
        "tailored to the context and audience. It often involves avoiding overly casual language, slang, or "
        "colloquialisms, and instead using clear, concise, and respectful language."
    ),
    grading_prompt=(
        "Professionalism: If the answer is written using a professional tone, below are the details for different scores: "
        "- Score 0: Language is extremely casual, informal, and may include slang or colloquialisms. Not suitable for "

zoltanctoth

long_text = 3000 * "Lorem ipsum dolor sit amet, consectetur adipiscing elit. "
COMPREHEND_LIMIT = 5000

lines = long_text.split(".")
current_text = ""
for line in lines:
    if len(current_text + line) > COMPREHEND_LIMIT:
        # EXECUTE COMPREHEND
        print(f"Executing Comprehend on {len(current_text)} characters")
        current_text = ""

zoltanctoth

import os

import mlflow
import uvicorn
from fastapi import FastAPI, HTTPException


class MLService:
    def __init__(self, model):
        self.model = model

zoltanctoth

models:
   - name: dim_listings_w_hosts
    tests:
      - dbt_expectations.expect_table_row_count_to_equal_other_table:
          compare_model: source('airbnb', 'listings')
    columns:
      - name: price
        tests:
          - dbt_expectations.expect_column_values_to_be_of_type:
              column_type: number

zoltanctoth

{% test valid_age(model, column_name) %}
SELECT COUNT(*) FROM {{ model }} 
   WHERE NOT {{ column_name}} BETWEEN 0 and 125
{% endtest %}

zoltanctoth

packages:
  - package: calogica/dbt_expectations
    version: [">=0.6.0", "<0.7.0"]

zoltanctoth

# Print a string without adding a newline
print("Hey, Python prints without a newline.", end ="") 

# Alternative solution
import sys
sys.stdout.write("Hey, Python prints without a newline.")

# You are part of an experiment on how well gists can be used as "StackOverflow". 
# Please add a comment or a star if you found this useful. :) Thanks!