2D MLP Assembly Tutorial Animation
Budget: $30 – $250 USD
I want to create a clear, step-by-step 2D animation that teaches the basics of building and training a simple multilayer perceptron for a toy prediction task. The viewer should see every stage unfold visually: individual neurons receiving inputs, weights and bias being applied, the weighted sum rolling into an activation function, the forward-propagated output compared with the expected value, and the error feeding back to correct the weights until the network converges on the right answer.
Because I prefer a text-only presentation, please rely on concise on-screen captions or animated callouts instead of a voice-over. Think clean, modern flat-style graphics rather than a whiteboard sketch or 3D render. The pace should feel like a live walkthrough: each scene pauses just long enough for the caption to be read before the next computational step animates in. The overall tone is educational but friendly, avoiding heavy mathematical jargon while still naming the essential concepts (sum, bias, activation, loss, update).
Deliverables
• 1080p MP4 (roughly 2–3 minutes)
• Layered project/source files so I can tweak text later
• Storyboard or animatic for approval before full production
Acceptance is straightforward: if a beginner can watch the finished video and verbally describe how an MLP forward-propagates, measures error, and updates weights until the output matches the target, we’re done.
Because I prefer a text-only presentation, please rely on concise on-screen captions or animated callouts instead of a voice-over. Think clean, modern flat-style graphics rather than a whiteboard sketch or 3D render. The pace should feel like a live walkthrough: each scene pauses just long enough for the caption to be read before the next computational step animates in. The overall tone is educational but friendly, avoiding heavy mathematical jargon while still naming the essential concepts (sum, bias, activation, loss, update).
Deliverables
• 1080p MP4 (roughly 2–3 minutes)
• Layered project/source files so I can tweak text later
• Storyboard or animatic for approval before full production
Acceptance is straightforward: if a beginner can watch the finished video and verbally describe how an MLP forward-propagates, measures error, and updates weights until the output matches the target, we’re done.