Upload README.md with huggingface_hub

README.md

@@ -4,25 +4,25 @@ tags:
 - emotion-recognition
 - audio-visual
 - multimodal
-- cremad
+- meld
 license: apache-2.0
 ---
 
-# AVERFormer-v4 (CREMA-D)
+# AVERFormer-v4 (MELD)
 
-Multimodal Audio-Visual-Text Emotion Recognition transformer trained on CREMA-D.
+Multimodal Audio-Visual-Text Emotion Recognition transformer trained on MELD.
 Code: https://github.com/mhussainahmad/AVERFormer
 
 ## Reported numbers
-- Best single-seed val wF1: **0.8320558803202199** (seed 100, epoch 12)
-- Best ensemble wF1: **0.8278 (top-3 ensemble)**
+- Best single-seed val wF1: **0.3815709082075283** (seed 1337, epoch 4)
+- Best ensemble wF1: **0.6150 (ensemble), 0.6128 (10-seed avg)**
 
-## Classes (6)
-['anger', 'disgust', 'fear', 'happy', 'neutral', 'sad']
+## Classes (7)
+['neutral', 'joy', 'sadness', 'anger', 'fear', 'disgust', 'surprise']
 
 ## Architecture
 - Audio: `microsoft/wavlm-large`  (16 kHz mono waveform)
-- Video: `MCG-NJU/videomae-large` (16 frames @ 224×224 RGB)
+- Video: `MCG-NJU/videomae-large` (16 frames @ 224x224 RGB)
 - Text:  `microsoft/deberta-v3-large` (speaker-aware ctx encoder)
 - Fusion: 2-layer cross-modal transformer, dim=512, 8 heads
 - Heads: face / voice / text / joint (all share class count)
@@ -34,8 +34,8 @@ import json, torch
 from huggingface_hub import hf_hub_download
 from models.averformer_v4 import AVERFormerV4
 
-cfg = json.load(open(hf_hub_download(repo_id="mhussainahmad/averformer-cremad-v4", filename="config.json")))
-ckpt = hf_hub_download(repo_id="mhussainahmad/averformer-cremad-v4", filename="pytorch_model.pth")
+cfg = json.load(open(hf_hub_download(repo_id="mhussainahmad/averformer-meld-v4", filename="config.json")))
+ckpt = hf_hub_download(repo_id="mhussainahmad/averformer-meld-v4", filename="pytorch_model.pth")
 
 model = AVERFormerV4(
     audio_backbone=cfg["audio_backbone"],
@@ -54,23 +54,7 @@ model.eval()
 ## Live inference
 
 ```bash
-python live_emotion_v4.py --repo_id mhussainahmad/averformer-cremad-v4
+python live_emotion_v4.py --repo_id mhussainahmad/averformer-meld-v4
 ```
 
 See `LIVE_INFERENCE_README.md` in the GitHub repo for full setup.
-
-## Training command (reference)
-
-```
-python train_v5_spec.py --corpus CREMA-D --classes 6 \
-  --audio microsoft/wavlm-large \
-  --video MCG-NJU/videomae-large \
-  --text  microsoft/deberta-v3-large \
-  --lora_r 16 \
-  --epochs 20 --bs 4 \
-  --grad_accum 4 \
-  --lr_head 0.0001 --lr_backbone 2e-05 \
-  --loss ce --class_weights sqrt-inv \
-  --select_metric wF1 \
-  --seed 100
-```