Instructions to use thomasjvu/lisper-gemma4-e2b-audio-lora with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- PEFT
How to use thomasjvu/lisper-gemma4-e2b-audio-lora with PEFT:
from peft import PeftModel from transformers import AutoModelForCausalLM base_model = AutoModelForCausalLM.from_pretrained("unsloth/gemma-4-e2b-it-unsloth-bnb-4bit") model = PeftModel.from_pretrained(base_model, "thomasjvu/lisper-gemma4-e2b-audio-lora") - Notebooks
- Google Colab
- Kaggle
- Local Apps
- Unsloth Studio
How to use thomasjvu/lisper-gemma4-e2b-audio-lora with Unsloth Studio:
Install Unsloth Studio (macOS, Linux, WSL)
curl -fsSL https://unsloth.ai/install.sh | sh # Run unsloth studio unsloth studio -H 0.0.0.0 -p 8888 # Then open http://localhost:8888 in your browser # Search for thomasjvu/lisper-gemma4-e2b-audio-lora to start chatting
Install Unsloth Studio (Windows)
irm https://unsloth.ai/install.ps1 | iex # Run unsloth studio unsloth studio -H 0.0.0.0 -p 8888 # Then open http://localhost:8888 in your browser # Search for thomasjvu/lisper-gemma4-e2b-audio-lora to start chatting
Using HuggingFace Spaces for Unsloth
# No setup required # Open https://huggingface.co/spaces/unsloth/studio in your browser # Search for thomasjvu/lisper-gemma4-e2b-audio-lora to start chatting
Load model with FastModel
pip install unsloth from unsloth import FastModel model, tokenizer = FastModel.from_pretrained( model_name="thomasjvu/lisper-gemma4-e2b-audio-lora", max_seq_length=2048, )
File size: 1,253 Bytes
4f75070 04004f8 | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 | {
"evaluation_name": "v18_hybrid_acoustic_gemma_heldout",
"status": "pass",
"count": 2000,
"success_count": 2000,
"effective_success_count": 2000,
"error_count": 0,
"hard_error_count": 0,
"hard_error_ids": [],
"truncated_count": 225,
"in_memory_retry_count": 75,
"acoustic_hint_count": 2000,
"acoustic_hint_match": 0.976,
"response_repaired_count": 2000,
"generation_fallback_count": 75,
"class_match": 0.976,
"class_match_successful_only": 0.976,
"clear_match": 0.989,
"clear_match_successful_only": 0.989,
"has_reason": 1.0,
"has_reason_successful_only": 1.0,
"has_corrective_cue": 1.0,
"has_corrective_cue_successful_only": 1.0,
"has_encouragement": 1.0,
"has_encouragement_successful_only": 1.0,
"format_exact": 1.0,
"format_exact_successful_only": 1.0,
"format_four_lines": 1.0,
"format_four_lines_successful_only": 1.0,
"detected_class_in_schema": 1.0,
"detected_class_in_schema_successful_only": 1.0,
"notes": [
"This is the v18 hybrid acoustic+Gemma held-out evaluation.",
"The lisp-class hint comes from acoustic features; Gemma generates the structured coaching response.",
"Do not interpret these metrics as a pure direct-Gemma raw-audio classification result."
]
}
|