Scene Perception Module 1
Multi-task transformer (distilroberta-base) for single-scene film analysis.
Produces 256-d scene embeddings consumed by Module 2.
11 Scene-Level Heads
| # | Head | Type | Output |
|---|---|---|---|
| 1 | emotional_valence | 4-class | Positive_Uplifting, Neutral_Complex, Tension_Action, Negative_Distressing |
| 2 | conflict_nature | 6-class | Physical_Danger, Psychological_Tension, Interpersonal_Conflict, Moral_Dilemma, Environmental_Threat, Unknown_Threat |
| 3 | acoustic_space | 6-class | Interior_Small, Interior_Large, Outdoor_Natural, Outdoor_Urban, Vehicle, Abstract |
| 4 | reality_layer | 5-class | Present, Memory, Dream, Internal, Distorted |
| 5 | score_dynamic_shape | 4-class | Build_Release, Sustained, Sudden_Drop, Flat |
| 6 | scene_interaction_tone | 5-class | Conflict, Bonding, Expository, Negotiation, Reflective |
| 7 | pacing_intensity | regression | 1โ10 |
| 8 | action_intensity | regression | 0โ10 |
| 9 | scene_tension_raw | regression | 1โ10 |
| 10 | scene_arousal | regression | 0โ1 |
| 11 | emotion_tags | multi-label 7 | Anger, Joy, Sadness, Fear, Disgust, Surprise, Neutral |
Inference Providers NEW
This model isn't deployed by any Inference Provider. ๐ Ask for provider support