Hugging Face's logo

barakplasma
/
translategemma-4b-it-android-task-quantized

Translation
LiteRT
LiteRT
android
on-device
gemma3
google-ai-edge
Model card Files
xet
 Community
1
barakplasma commited on
Commit
c4b7dc7
·
unverified ·
1 Parent(s): 1fb53c0

Upload logs/convert_none.log with huggingface_hub

Browse files
Files changed (1) hide show
  1. logs/convert_none.log +60 -0
logs/convert_none.log ADDED
@@ -0,0 +1,60 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ [!] HF_TOKEN is not set. If model is public, you can pass --allow-no-token.
2
+ [+] Using existing model dir: /home/ubuntu/translategemma-4b-it
3
+ [+] ARCH: {'model_type': 'gemma3', 'architecture': 'Gemma3ForConditionalGeneration', 'vocab_size': 262208}
4
+ [+] Strategy 1: litert-torch native
5
+ [+] Gemma3 builders available: ['build_model_1b', 'build_model_270m']
6
+ [+] Trying litert_torch.generative.examples.gemma3.gemma3.build_model_1b ...
7
+ [!] failed: 'NoneType' object has no attribute 'eval'
8
+ [+] Trying litert_torch.generative.examples.gemma3.gemma3.build_model_270m ...
9
+ [!] failed: 'NoneType' object has no attribute 'eval'
10
+ [!] Strategy 1 did not find a compatible builder
11
+ [+] Strategy 2: ai_edge_torch generic (wrapped logits-only)
12
+ [+] Loading HF model on CPU with dtype=torch.float32 ...
13
+ `torch_dtype` is deprecated! Use `dtype` instead!
14
+
15
+ Loading weights: 0%| | 0/883 [00:00<?, ?it/s]
16
+ Loading weights: 8%|████████ | 67/883 [00:00<00:01, 498.07it/s]
17
+ Loading weights: 13%|██████████████ | 117/883 [00:00<00:01, 440.12it/s]
18
+ Loading weights: 18%|███████████████████▎ | 161/883 [00:00<00:02, 330.40it/s]
19
+ Loading weights: 22%|███████████████████████▊ | 198/883 [00:00<00:02, 337.14it/s]
20
+ Loading weights: 27%|████████████████████████████▍ | 237/883 [00:00<00:01, 340.87it/s]
21
+ Loading weights: 31%|████████████████████████████████▋ | 272/883 [00:00<00:01, 340.67it/s]
22
+ Loading weights: 35%|████████████████████████████████████▊ | 307/883 [00:00<00:01, 342.47it/s]
23
+ Loading weights: 39%|█████████████████████████████████████████ | 342/883 [00:00<00:01, 334.91it/s]
24
+ Loading weights: 43%|█████████████████████████████████████████████▏ | 376/883 [00:01<00:01, 288.99it/s]
25
+ Loading weights: 47%|█████████████████████████████████████████████████▍ | 412/883 [00:01<00:01, 307.09it/s]
26
+ Loading weights: 54%|████████████████████████████████████████████████████████▊ | 473/883 [00:01<00:01, 388.67it/s]
27
+ Loading weights: 86%|██████████████████████████████████████████████████████████████████████████████████████████▏ | 758/883 [00:01<00:00, 1076.01it/s]
28
+ Loading weights: 100%|██████████████████████████████████████████████████████████████████████████████████████████████████████████| 883/883 [00:01<00:00, 597.78it/s]
29
+ (00:00) [START] LiteRT-Torch Convert
30
+ (00:00) [START] LiteRT-Torch Convert > Torch Export: serving_default
31
+ (00:08) [START] LiteRT-Torch Convert > Torch Export: serving_default > ExportedProgram Run Decompositions
32
+ (00:25) [ DONE] LiteRT-Torch Convert > Torch Export: serving_default > ExportedProgram Run Decompositions (+00:16)
33
+ (00:25) [ DONE] LiteRT-Torch Convert > Torch Export: serving_default (+00:25)
34
+ (00:25) [START] LiteRT-Torch Convert > Run FX Passes
35
+ (00:26) [START] LiteRT-Torch Convert > Run FX Passes > ExportedProgram Run Decompositions
36
+ (00:26) [ DONE] LiteRT-Torch Convert > Run FX Passes > ExportedProgram Run Decompositions (+00:00)
37
+ (00:26) [ DONE] LiteRT-Torch Convert > Run FX Passes (+00:01)
38
+ (00:26) [START] LiteRT-Torch Convert > Lower to MLIR: serving_default
39
+ (00:26) [START] LiteRT-Torch Convert > Lower to MLIR: serving_default > ExportedProgram Run Decompositions
40
+ (00:42) [ DONE] LiteRT-Torch Convert > Lower to MLIR: serving_default > ExportedProgram Run Decompositions (+00:15)
41
+ (00:42) [START] LiteRT-Torch Convert > Lower to MLIR: serving_default > ExportedProgram Run Decompositions
42
+ (00:59) [ DONE] LiteRT-Torch Convert > Lower to MLIR: serving_default > ExportedProgram Run Decompositions (+00:16)
43
+ (00:59) [START] LiteRT-Torch Convert > Lower to MLIR: serving_default > Create MLIR Module
44
+ WARNING:jax._src.xla_bridge:An NVIDIA GPU may be present on this machine, but a CUDA-enabled jaxlib is not installed. Falling back to cpu.
45
+ (01:32) [ DONE] LiteRT-Torch Convert > Lower to MLIR: serving_default > Create MLIR Module (+00:32)
46
+ (01:32) [ DONE] LiteRT-Torch Convert > Lower to MLIR: serving_default (+01:05)
47
+ (01:32) [START] LiteRT-Torch Convert > Merge MLIR Modules
48
+ (01:32) [ DONE] LiteRT-Torch Convert > Merge MLIR Modules (+00:00)
49
+ (01:32) [START] LiteRT-Torch Convert > Run LiteRT Converter Passes
50
+ (07:03) [ DONE] LiteRT-Torch Convert > Run LiteRT Converter Passes (+05:31)
51
+ (07:03) [ DONE] LiteRT-Torch Convert (+07:03)
52
+ (00:00) [START] Write Model to /home/ubuntu/tflite_output/none/translategemma-4b-it-generic-none.tflite
53
+ Module size is greater than 2GB
54
+ (00:07) [ DONE] Write Model to /home/ubuntu/tflite_output/none/translategemma-4b-it-generic-none.tflite (+00:07)
55
+ [+] Strategy 2 success: /home/ubuntu/tflite_output/none/translategemma-4b-it-generic-none.tflite
56
+ [!] Generic TFLite may not have MediaPipe LLM prefill/decode signatures.
57
+ [+] Bundling .task -> /home/ubuntu/output/translategemma-4b-it-none.task
58
+ [+] BundleConfig params: ['tflite_model', 'tokenizer_model', 'start_token', 'stop_tokens', 'output_filename', 'enable_bytes_to_unicode_mapping', 'system_prompt', 'prompt_prefix_user', 'prompt_suffix_user', 'prompt_prefix_model', 'prompt_suffix_model', 'prompt_prefix_system', 'prompt_suffix_system', 'user_role_token', 'system_role_token', 'model_role_token', 'end_role_token', 'tflite_embedder', 'tflite_per_layer_embedder', 'tflite_vision_encoder', 'tflite_vision_adapter']
59
+ [+] DONE: /home/ubuntu/output/translategemma-4b-it-none.task
60
+ [+] Size: 14810.01 MB