Spaces:
Sleeping
Sleeping
Update app.py
Browse files
app.py
CHANGED
|
@@ -1,4 +1,4 @@
|
|
| 1 |
-
# app.py — Íslenskt ASR – ZeroGPU
|
| 2 |
import os
|
| 3 |
os.environ["OMP_NUM_THREADS"] = "1"
|
| 4 |
os.environ["PYTORCH_CUDA_ALLOC_CONF"] = "garbage_collection_threshold:0.6,max_split_size_mb:128"
|
|
@@ -9,20 +9,17 @@ from transformers import pipeline
|
|
| 9 |
import torch
|
| 10 |
import gc
|
| 11 |
|
| 12 |
-
# ——————————————————————————————
|
| 13 |
-
# Global pipeline – lazy load + auto-rebuild on OOM
|
| 14 |
-
# ——————————————————————————————
|
| 15 |
MODEL_NAME = "palli23/whisper-small-sam_spjall"
|
| 16 |
-
pipe = None
|
| 17 |
|
| 18 |
@spaces.GPU(duration=180, max_batch_size=4)
|
| 19 |
def get_or_refresh_pipeline():
|
| 20 |
global pipe
|
| 21 |
|
| 22 |
-
#
|
| 23 |
if pipe is not None:
|
| 24 |
try:
|
| 25 |
-
_ = pipe.model.device
|
| 26 |
except Exception:
|
| 27 |
print("GPU context lost → rebuilding pipeline...")
|
| 28 |
pipe = None
|
|
@@ -42,17 +39,16 @@ def get_or_refresh_pipeline():
|
|
| 42 |
|
| 43 |
return pipe
|
| 44 |
|
| 45 |
-
# ——————————————————————————————
|
| 46 |
-
# Transcription – memory-safe + fixed del bug
|
| 47 |
-
# ——————————————————————————————
|
| 48 |
def transcribe_3min(audio_path):
|
| 49 |
if not audio_path:
|
| 50 |
return "Hlaðið upp hljóðskrá (mp3/wav, max 5 mín)"
|
| 51 |
|
|
|
|
|
|
|
| 52 |
try:
|
| 53 |
-
|
| 54 |
|
| 55 |
-
result =
|
| 56 |
audio_path,
|
| 57 |
chunk_length_s=30,
|
| 58 |
stride_length_s=(6, 0),
|
|
@@ -63,33 +59,29 @@ def transcribe_3min(audio_path):
|
|
| 63 |
|
| 64 |
text = result["text"].strip()
|
| 65 |
|
| 66 |
-
#
|
| 67 |
if "chunks" in result:
|
| 68 |
del result["chunks"]
|
| 69 |
|
| 70 |
-
# Aggressive cleanup
|
| 71 |
gc.collect()
|
| 72 |
torch.cuda.empty_cache()
|
| 73 |
|
| 74 |
return text if text else "(ekkert tal greint)"
|
| 75 |
|
| 76 |
except torch.cuda.OutOfMemoryError:
|
| 77 |
-
print("OOM → forcing
|
| 78 |
-
global
|
| 79 |
-
pipe = None
|
| 80 |
gc.collect()
|
| 81 |
torch.cuda.empty_cache()
|
| 82 |
return "Of mikið minni notað – bíddu 10 sek og prófaðu aftur (ZeroGPU takmörk)"
|
| 83 |
|
| 84 |
except Exception as e:
|
| 85 |
-
return f"
|
| 86 |
|
| 87 |
-
#
|
| 88 |
-
# UI
|
| 89 |
-
# ——————————————————————————————
|
| 90 |
with gr.Blocks(title="Íslenskt ASR") as demo:
|
| 91 |
gr.Markdown("# Íslenskt ASR – 3–5 mín hljóð")
|
| 92 |
-
gr.Markdown("**Whisper-small fínstillt á íslensku · mjög lágur WER**")
|
| 93 |
gr.Markdown("**Hafa samband:** pallinr1@protonmail.com")
|
| 94 |
gr.Markdown("> Keyrt á **ZeroGPU** – fyrsta ræsing tekur 15–30 sek, síðan hröð")
|
| 95 |
|
|
@@ -110,9 +102,7 @@ with gr.Blocks(title="Íslenskt ASR") as demo:
|
|
| 110 |
- Ef þú færð minnisvillu → bíddu öðruhvolf og prófaðu aftur
|
| 111 |
""")
|
| 112 |
|
| 113 |
-
#
|
| 114 |
-
# Launch
|
| 115 |
-
# ——————————————————————————————
|
| 116 |
demo.launch(
|
| 117 |
auth=None,
|
| 118 |
share=True,
|
|
|
|
| 1 |
+
# app.py — Íslenskt ASR – ZeroGPU 100% stable (Dec 2025 final fixed version)
|
| 2 |
import os
|
| 3 |
os.environ["OMP_NUM_THREADS"] = "1"
|
| 4 |
os.environ["PYTORCH_CUDA_ALLOC_CONF"] = "garbage_collection_threshold:0.6,max_split_size_mb:128"
|
|
|
|
| 9 |
import torch
|
| 10 |
import gc
|
| 11 |
|
|
|
|
|
|
|
|
|
|
| 12 |
MODEL_NAME = "palli23/whisper-small-sam_spjall"
|
| 13 |
+
pipe = None # Global pipeline
|
| 14 |
|
| 15 |
@spaces.GPU(duration=180, max_batch_size=4)
|
| 16 |
def get_or_refresh_pipeline():
|
| 17 |
global pipe
|
| 18 |
|
| 19 |
+
# Check if pipeline is broken or GPU context died
|
| 20 |
if pipe is not None:
|
| 21 |
try:
|
| 22 |
+
_ = pipe.model.device # Quick health check
|
| 23 |
except Exception:
|
| 24 |
print("GPU context lost → rebuilding pipeline...")
|
| 25 |
pipe = None
|
|
|
|
| 39 |
|
| 40 |
return pipe
|
| 41 |
|
|
|
|
|
|
|
|
|
|
| 42 |
def transcribe_3min(audio_path):
|
| 43 |
if not audio_path:
|
| 44 |
return "Hlaðið upp hljóðskrá (mp3/wav, max 5 mín)"
|
| 45 |
|
| 46 |
+
global pipe # ← Now declared at the very top of the function → no error!
|
| 47 |
+
|
| 48 |
try:
|
| 49 |
+
current_pipe = get_or_refresh_pipeline()
|
| 50 |
|
| 51 |
+
result = current_pipe(
|
| 52 |
audio_path,
|
| 53 |
chunk_length_s=30,
|
| 54 |
stride_length_s=(6, 0),
|
|
|
|
| 59 |
|
| 60 |
text = result["text"].strip()
|
| 61 |
|
| 62 |
+
# Clean up chunks if present
|
| 63 |
if "chunks" in result:
|
| 64 |
del result["chunks"]
|
| 65 |
|
|
|
|
| 66 |
gc.collect()
|
| 67 |
torch.cuda.empty_cache()
|
| 68 |
|
| 69 |
return text if text else "(ekkert tal greint)"
|
| 70 |
|
| 71 |
except torch.cuda.OutOfMemoryError:
|
| 72 |
+
print("OOM detected → forcing full pipeline reload")
|
| 73 |
+
pipe = None # This is now allowed because global declared first
|
|
|
|
| 74 |
gc.collect()
|
| 75 |
torch.cuda.empty_cache()
|
| 76 |
return "Of mikið minni notað – bíddu 10 sek og prófaðu aftur (ZeroGPU takmörk)"
|
| 77 |
|
| 78 |
except Exception as e:
|
| 79 |
+
return f"Óvænt villa: {str(e)}"
|
| 80 |
|
| 81 |
+
# ————————————————————— UI —————————————————————
|
|
|
|
|
|
|
| 82 |
with gr.Blocks(title="Íslenskt ASR") as demo:
|
| 83 |
gr.Markdown("# Íslenskt ASR – 3–5 mín hljóð")
|
| 84 |
+
gr.Markdown("**Whisper-small fínstillt á íslensku spjalli · mjög lágur WER**")
|
| 85 |
gr.Markdown("**Hafa samband:** pallinr1@protonmail.com")
|
| 86 |
gr.Markdown("> Keyrt á **ZeroGPU** – fyrsta ræsing tekur 15–30 sek, síðan hröð")
|
| 87 |
|
|
|
|
| 102 |
- Ef þú færð minnisvillu → bíddu öðruhvolf og prófaðu aftur
|
| 103 |
""")
|
| 104 |
|
| 105 |
+
# ————————————————————— Launch —————————————————————
|
|
|
|
|
|
|
| 106 |
demo.launch(
|
| 107 |
auth=None,
|
| 108 |
share=True,
|