Upload 8 files

README.md

@@ -0,0 +1,63 @@
+
+# SpotSound: Enhancing Large Audio-Language Models with Fine-Grained Temporal Grounding
+
+[![GitHub](https://img.shields.io/badge/GitHub-Repo-black?logo=github)](https://github.com/LoieSun/SpotSound) 
+[![Paper](https://img.shields.io/badge/arXiv-Paper-red?logo=arxiv)](#) 
+[![Benchmark](https://img.shields.io/badge/🤗_HuggingFace-Benchmark-yellow)](https://huggingface.co/datasets/Loie/SpotSound-Bench) 
+
+## Model Summary
+
+**SpotSound** is a model designed to enhance Large Audio-Language Models (ALMs) with fine-grained temporal grounding capabilities. Built on top of [Audio Flamingo 3](https://huggingface.co/nvidia/audio-flamingo-3), SpotSound is capable of accurately pinpointing the exact start and end timestamps of specific acoustic events within long, untrimmed audio recordings based on natural language queries.
+
+This model is particularly effective for "needle-in-a-haystack" audio retrieval tasks, where short target sounds are embedded within complex background noise.
+
+
+## Usage / Quick Start
+
+To use SpotSound for inference, you need to download both the base **Audio Flamingo 3** model and the **SpotSound** checkpoint.
+
+### 1. Installation
+
+First, clone the official [SpotSound GitHub repository](#) and set up the environment:
+
+```bash 
+conda create -n SpotSound python=3.10
+conda activate SpotSound
+pip install torch==2.1.2 torchvision==0.16.2 torchaudio==2.1.2 --index-url https://download.pytorch.org/whl/cu121
+pip install -r requirements.txt
+```
+
+### 2. Run Inference
+
+You can run inference directly from the command line using the provided script in the GitHub repository. Specify the path to the downloaded Audio Flamingo 3 model, your SpotSound checkpoint, the target audio file, and your text query.
+
+```bash
+export CUDA_VISIBLE_DEVICES=0 
+python inference.py \
+    --pretrain_model path_to_audioflamingo3 \
+    --checkpoint ckpt/spotsound \
+    --audio_path data/audio.wav \
+    --query "dog barking"
+```
+
+## Citation
+
+If you use SpotSound or our benchmark in your research, please cite our paper:
+
+```bibtex
+@inproceedings{sun2026spotsound,
+    title={SpotSound: Enhancing Large Audio-Language Models with Fine-Grained Temporal Grounding},
+    author={Sun, Luoyi and Zhou, Xiao and Li, Zeqian and Zhang, Ya and Wang, Yanking and Xie, Weidi},
+    year={2026}
+}
+```
+
+## Acknowledgements
+
+This project builds upon several excellent open-source efforts, notably:
+- **[Audio Flamingo 3](https://github.com/NVIDIA/audio-flamingo/tree/audio_flamingo_3)** by NVIDIA.
+- **[UniTime](https://github.com/Lzq5/UniTime)** for temporal grounding insights.
+
+## Contact
+
+For any questions or issues, please contact: loiesun411@gmail.com or open an issue on our GitHub repository.

adapter_config.json

@@ -0,0 +1,127 @@
+{
+  "alora_invocation_tokens": null,
+  "alpha_pattern": {},
+  "arrow_config": null,
+  "auto_mapping": null,
+  "base_model_name_or_path": "audio-flamingo-3-hf",
+  "bias": "none",
+  "corda_config": null,
+  "ensure_weight_tying": false,
+  "eva_config": null,
+  "exclude_modules": null,
+  "fan_in_fan_out": false,
+  "inference_mode": true,
+  "init_lora_weights": true,
+  "layer_replication": null,
+  "layers_pattern": null,
+  "layers_to_transform": null,
+  "loftq_config": {},
+  "lora_alpha": 16,
+  "lora_bias": false,
+  "lora_dropout": 0.1,
+  "megatron_config": null,
+  "megatron_core": "megatron.core",
+  "modules_to_save": null,
+  "peft_type": "LORA",
+  "peft_version": "0.18.1",
+  "qalora_group_size": 16,
+  "r": 8,
+  "rank_pattern": {},
+  "revision": null,
+  "target_modules": [
+    "model.layers.0.self_attn.k_proj",
+    "model.layers.17.self_attn.k_proj",
+    "model.layers.24.self_attn.q_proj",
+    "model.layers.2.self_attn.q_proj",
+    "model.layers.0.self_attn.v_proj",
+    "model.layers.7.self_attn.k_proj",
+    "model.layers.15.self_attn.v_proj",
+    "model.layers.19.self_attn.k_proj",
+    "down_proj",
+    "model.layers.12.self_attn.v_proj",
+    "model.layers.21.self_attn.v_proj",
+    "model.layers.23.self_attn.v_proj",
+    "model.layers.26.self_attn.k_proj",
+    "model.layers.19.self_attn.v_proj",
+    "model.layers.15.self_attn.k_proj",
+    "model.layers.24.self_attn.k_proj",
+    "model.layers.18.self_attn.q_proj",
+    "model.layers.20.self_attn.k_proj",
+    "model.layers.5.self_attn.v_proj",
+    "model.layers.10.self_attn.v_proj",
+    "model.layers.22.self_attn.k_proj",
+    "model.layers.3.self_attn.k_proj",
+    "model.layers.6.self_attn.k_proj",
+    "model.layers.9.self_attn.k_proj",
+    "model.layers.22.self_attn.q_proj",
+    "model.layers.5.self_attn.k_proj",
+    "model.layers.25.self_attn.v_proj",
+    "model.layers.23.self_attn.q_proj",
+    "model.layers.13.self_attn.v_proj",
+    "model.layers.3.self_attn.q_proj",
+    "model.layers.10.self_attn.q_proj",
+    "model.layers.13.self_attn.k_proj",
+    "model.layers.23.self_attn.k_proj",
+    "model.layers.15.self_attn.q_proj",
+    "model.layers.1.self_attn.k_proj",
+    "model.layers.2.self_attn.v_proj",
+    "gate_proj",
+    "model.layers.17.self_attn.v_proj",
+    "model.layers.21.self_attn.q_proj",
+    "model.layers.27.self_attn.q_proj",
+    "model.layers.7.self_attn.v_proj",
+    "model.layers.11.self_attn.q_proj",
+    "model.layers.16.self_attn.v_proj",
+    "model.layers.11.self_attn.v_proj",
+    "model.layers.12.self_attn.q_proj",
+    "model.layers.18.self_attn.v_proj",
+    "model.layers.3.self_attn.v_proj",
+    "model.layers.27.self_attn.v_proj",
+    "model.layers.26.self_attn.v_proj",
+    "model.layers.21.self_attn.k_proj",
+    "model.layers.8.self_attn.q_proj",
+    "model.layers.25.self_attn.q_proj",
+    "model.layers.5.self_attn.q_proj",
+    "model.layers.9.self_attn.q_proj",
+    "model.layers.18.self_attn.k_proj",
+    "model.layers.4.self_attn.k_proj",
+    "model.layers.0.self_attn.q_proj",
+    "model.layers.11.self_attn.k_proj",
+    "model.layers.4.self_attn.v_proj",
+    "model.layers.20.self_attn.q_proj",
+    "model.layers.10.self_attn.k_proj",
+    "model.layers.22.self_attn.v_proj",
+    "model.layers.14.self_attn.k_proj",
+    "model.layers.1.self_attn.q_proj",
+    "model.layers.12.self_attn.k_proj",
+    "model.layers.13.self_attn.q_proj",
+    "model.layers.20.self_attn.v_proj",
+    "up_proj",
+    "model.layers.25.self_attn.k_proj",
+    "model.layers.17.self_attn.q_proj",
+    "model.layers.1.self_attn.v_proj",
+    "model.layers.4.self_attn.q_proj",
+    "model.layers.19.self_attn.q_proj",
+    "model.layers.14.self_attn.v_proj",
+    "model.layers.27.self_attn.k_proj",
+    "model.layers.14.self_attn.q_proj",
+    "model.layers.7.self_attn.q_proj",
+    "model.layers.16.self_attn.q_proj",
+    "o_proj",
+    "model.layers.16.self_attn.k_proj",
+    "model.layers.6.self_attn.v_proj",
+    "model.layers.8.self_attn.k_proj",
+    "model.layers.2.self_attn.k_proj",
+    "model.layers.26.self_attn.q_proj",
+    "model.layers.24.self_attn.v_proj",
+    "model.layers.8.self_attn.v_proj",
+    "model.layers.9.self_attn.v_proj",
+    "model.layers.6.self_attn.q_proj"
+  ],
+  "target_parameters": null,
+  "task_type": "CAUSAL_LM",
+  "trainable_token_indices": null,
+  "use_dora": false,
+  "use_qalora": false,
+  "use_rslora": false
+}

adapter_model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ab87de32fc6ae5acfaccc22c5f9792230a4ac8f11ca15f968d21b9c26075cf5e
+size 80797976

optimizer.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c6b92c23c4174f5dabf3c48a10081edd82868344e2cf75016708d02fa257ac2b
+size 161810282

rng_state.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:2ab158b3630065e0379f440b006083e3cb588fe2f21ffcb5a3a4b57b941537c1
+size 14244

scheduler.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:1996f8e7b4f0cb933e459b73d89902fdf29339d53a3bf845c48f923c01b102c0
+size 1064

training_args.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:0bc632d0ab1a2384fc16b375a26abd5fb27276c54e503f67ab4fa0533fc768cb
+size 4856