Chapter 4: full model

2025-12-13 11:26:26 -08:00
parent 2f3c40b162
commit e29b25c162
1 changed files with 243 additions and 48 deletions
--- a/crates/llmfs/src/command/sample_data.rs
+++ b/crates/llmfs/src/command/sample_data.rs
@@ -3,7 +3,10 @@ use burn::{
 	Tensor,
 	backend::{Cuda, cuda::CudaDevice},
 	module::{Module, Param, ParamId},
-	nn::{Dropout, Embedding, EmbeddingConfig},
+	nn::{
+		Dropout, Embedding, EmbeddingConfig, LayerNorm, LayerNormConfig,
+		transformer::{PositionWiseFeedForward, PositionWiseFeedForwardConfig},
+	},
 	prelude::Backend,
 	tensor::{Bool, Distribution, Int, activation::softmax},
 };
@@ -35,26 +38,54 @@ pub struct SampleDataArgs {
 	skip: usize,
 }

+#[derive(Debug, Clone)]
+pub struct Config {
+	/// Number of tokens
+	pub vocab_size: u32,
+
+	/// Maximum number of input tokens with positional embeddings
+	pub context_size: usize,
+
+	/// Dimension of each token's embedding
+	pub embed_dim: usize,
+
+	/// Number of attention heads
+	pub n_heads: usize,
+
+	/// Dimension of each attn head
+	pub head_dim: usize,
+
+	/// Number of transformer blocks
+	pub n_layers: usize,
+
+	pub embed_drop_rate: f64,
+	pub attention_drop_rate: f64,
+	pub shortcut_drop_rate: f64,
+}
+
 impl SampleDataArgs {
 	pub fn run(self, _mp: Option<MultiProgress>) -> Result<()> {
 		let device = CudaDevice::new(0);

 		let iter = DataReader::new(1, &self.data_dir).context("while initializing data reader")?;

-		let tokenizer = File::open(self.tokenizer).context("while opening tokenizer")?;
+		let tokenizer = File::open(&self.tokenizer).context("while opening tokenizer")?;
 		let tokenizer: Tokenizer =
 			serde_json::from_reader(tokenizer).context("while loading tokenizer")?;

-		let context_size = 4;
-		let stride = 4;
+		let config = Config {
+			vocab_size: tokenizer.vocab_size(),
+			context_size: 4,
+			embed_dim: 768,
+			n_heads: 12,
+			head_dim: 64, // = 768 / 12
+			n_layers: 12,
+			embed_drop_rate: 0.1,
+			attention_drop_rate: 0.1,
+			shortcut_drop_rate: 0.1,
+		};

-		// Dimension of each token vector
-		let embedding_dim = 3;
-		// attention
-		let head_dim = 2;
-		let n_heads = 2;
-
-		let dropout = 0.5f64;
+		let stride = config.context_size;

 		let batch_size = 10;
 		let mut input_batch = Vec::with_capacity(batch_size);
@@ -63,42 +94,64 @@ impl SampleDataArgs {
 		#[expect(clippy::unwrap_used)] // Lazy error handling
 		let iter = iter.map(|x| x.unwrap()).skip(self.skip).take(self.n);

-		// TODO: what is this?
-		let tok_embedder = EmbeddingConfig::new(tokenizer.vocab_size() as usize, embedding_dim);
-		let tok_embedder: Embedding<Cuda> = tok_embedder.init(&device);
+		let model = GptModel::new(&config, &device);

-		// TODO: do we want to train this?
-		let pos_embedder = EmbeddingConfig::new(context_size, embedding_dim);
-		let pos_embedder: Embedding<Cuda> = pos_embedder.init(&device);
+		// Text generation routine

-		let pos_tensor: Tensor<Cuda, 2, Int> =
-			Tensor::arange(0..context_size as i64, &device).unsqueeze_dim(0);
+		/*
+		{
+			let init = "Initial context. This is ";
+			let tokens = tokenizer.encode(&init);

-		// [context_size, dim]
-		let pos_embedding = pos_embedder.forward(pos_tensor).squeeze_dim::<2>(0);
+			let n_tokens = tokens.len();
+			let input: Array1<u32> = Array1::from_vec(tokens);
+			let mut input: Tensor<Cuda, 1, Int> =
+				Tensor::<_, 1, Int>::from_ints(input.as_slice().unwrap(), &device)
+					.reshape([n_tokens]);

-		let attention: MultiheadAttention<Cuda> = MultiheadAttention::new(
-			embedding_dim,
-			head_dim,
-			n_heads,
-			context_size,
-			dropout,
-			&device,
-		);
+			for _ in 0..100 {
+				let tokens: Vec<u32> = input.clone().to_data().convert::<u32>().into_vec().unwrap();
+				println!("{:?}", tokens);
+				println!("{}", tokenizer.decode(&tokens));
+
+				// Crop idx to context size;
+				let batch = input
+					.clone()
+					.slice([0..config.context_size])
+					.unsqueeze_dim(0);
+
+				// shape: [tokens, vocab_size]
+				let logits = model.forward(batch).squeeze_dim::<2>(0);
+
+				// shape: [vocab_size]
+				let logits = logits.slice([config.context_size - 1]).squeeze_dim::<1>(0);
+
+				let probs = softmax(logits, 0); // shape: [n_tokens]
+				let id_next = probs.argmax(0); //  shape: [1]
+
+				input = Tensor::cat(vec![input.slice([1..]), id_next], 0);
+			}
+		}
+		*/

 		for i in iter {
 			let tokens = tokenizer.encode(&i);

-			for (a, b) in tokens
-				.windows(context_size)
-				.step_by(stride)
-				.zip(tokens[stride..].windows(context_size).step_by(stride))
-			{
+			// Skip small texts.
+			// TODO: do this better
+			// TODO: non-uniform batches?
+			if tokens.len() < config.context_size {
+				continue;
+			}
+
+			for (a, b) in tokens.windows(config.context_size).step_by(stride).zip(
+				tokens[stride..]
+					.windows(config.context_size)
+					.step_by(stride),
+			) {
 				input_batch.push(a.to_owned());
 				output_batch.push(b.to_owned());

-				// TODO: non-uniform batches?
-
 				/*
 				let context = a;
 				let desired = &b[b.len() - 1..];
@@ -109,12 +162,13 @@ impl SampleDataArgs {
 				println!("{input:?} -> {target:?}");
 				*/

-				// TODO: last batch
 				if input_batch.len() >= batch_size {
-					let shape = [batch_size, context_size];
+					let shape = [input_batch.len(), config.context_size];

 					let input = std::mem::replace(&mut input_batch, Vec::with_capacity(batch_size));
 					let input: Array2<u32> = Array2::from_shape_fn(shape, |(a, b)| input[a][b]);
+
+					#[expect(clippy::unwrap_used)]
 					let input: Tensor<Cuda, 2, Int> =
 						Tensor::<_, 1, Int>::from_ints(input.as_slice().unwrap(), &device)
 							.reshape(shape);
@@ -122,25 +176,44 @@ impl SampleDataArgs {
 					let output =
 						std::mem::replace(&mut output_batch, Vec::with_capacity(batch_size));
 					let output: Array2<u32> = Array2::from_shape_fn(shape, |(a, b)| output[a][b]);
+
+					#[expect(clippy::unwrap_used)]
 					let output: Tensor<Cuda, 2, Int> =
 						Tensor::<_, 1, Int>::from_ints(output.as_slice().unwrap(), &device)
 							.reshape(shape);

-					// Input token embeddings
-					// dim: [batch, token, dim]
-					let tok_e = tok_embedder.forward(input);
-					let tok_e = tok_e.add(pos_embedding.clone().unsqueeze_dim(0));
-
-					// Attention
-
-					// shape: [batch, tokens, out_dim]
-					let a = attention.forward(tok_e);
+					self.batch(&config, input, &model);
 				}
 			}
 		}

+		if !input_batch.is_empty() {
+			let shape = [input_batch.len(), config.context_size];
+
+			let input = std::mem::replace(&mut input_batch, Vec::with_capacity(batch_size));
+			let input: Array2<u32> = Array2::from_shape_fn(shape, |(a, b)| input[a][b]);
+
+			#[expect(clippy::unwrap_used)]
+			let input: Tensor<Cuda, 2, Int> =
+				Tensor::<_, 1, Int>::from_ints(input.as_slice().unwrap(), &device).reshape(shape);
+
+			let output = std::mem::replace(&mut output_batch, Vec::with_capacity(batch_size));
+			let output: Array2<u32> = Array2::from_shape_fn(shape, |(a, b)| output[a][b]);
+
+			#[expect(clippy::unwrap_used)]
+			let output: Tensor<Cuda, 2, Int> =
+				Tensor::<_, 1, Int>::from_ints(output.as_slice().unwrap(), &device).reshape(shape);
+
+			self.batch(&config, input, &model);
+		}
+
 		Ok(())
 	}
+
+	fn batch(&self, _cfg: &Config, input: Tensor<Cuda, 2, Int>, model: &GptModel<Cuda>) {
+		let logits = model.forward(input);
+		println!("{:?}", logits.shape());
+	}
 }

 /// Multihead attention.
@@ -247,7 +320,7 @@ impl<B: Backend> MultiheadAttention<B> {

 			dropout: Dropout { prob: dropout },

-			utri_mask: Tensor::<B, 2, _>::tril_mask([context_length, context_length], 0, &device),
+			utri_mask: Tensor::<B, 2, _>::tril_mask([context_length, context_length], 0, device),
 		}
 	}

@@ -348,3 +421,125 @@ impl<B: Backend> MultiheadAttention<B> {
 		return context_vec;
 	}
 }
+
+#[derive(Module, Debug)]
+pub struct GptModel<B: Backend> {
+	embedder_tok: Embedding<B>,
+	embedder_pos: Embedding<B>,
+	embedder_drop: Dropout,
+
+	trf_blocks: Vec<TransformerBlock<B>>,
+	final_norm: LayerNorm<B>,
+	out_head: Param<Tensor<B, 2>>,
+}
+
+impl<B: Backend> GptModel<B> {
+	pub fn new(cfg: &Config, device: &B::Device) -> Self {
+		let out_head_shape = [cfg.embed_dim, cfg.vocab_size as usize];
+
+		Self {
+			embedder_tok: EmbeddingConfig::new(cfg.vocab_size as usize, cfg.embed_dim).init(device),
+
+			embedder_pos: EmbeddingConfig::new(cfg.context_size, cfg.embed_dim).init(device),
+
+			embedder_drop: Dropout {
+				prob: cfg.embed_drop_rate,
+			},
+
+			trf_blocks: (0..cfg.n_layers)
+				.map(|_| TransformerBlock::new(cfg, device))
+				.collect(),
+
+			final_norm: LayerNormConfig::new(cfg.embed_dim).init(device),
+
+			out_head: Param::uninitialized(
+				ParamId::new(),
+				move |device, is_require_grad| {
+					Tensor::random(out_head_shape, Distribution::Default, device)
+						.set_require_grad(is_require_grad)
+				},
+				device.clone(),
+				true,
+				out_head_shape.into(),
+			),
+		}
+	}
+
+	pub fn forward(&self, input: Tensor<B, 2, Int>) -> Tensor<B, 3> {
+		let n_tokens = input.shape()[1];
+
+		let embed_tok = self.embedder_tok.forward(input.clone());
+		let embed_pos = self
+			.embedder_tok
+			.forward(Tensor::arange(0..n_tokens as i64, &input.device()).unsqueeze_dim(0));
+
+		let x = embed_tok + embed_pos;
+		let x = self.embedder_drop.forward(x);
+		let x = self.trf_blocks.iter().fold(x, |x, l| l.forward(x));
+		let x = self.final_norm.forward(x);
+		let logits = x.matmul(self.out_head.val().unsqueeze_dim(0));
+
+		return logits;
+	}
+}
+
+#[derive(Module, Debug)]
+pub struct TransformerBlock<B: Backend> {
+	attention: MultiheadAttention<B>,
+
+	/// TODO: wtf?
+	ff: PositionWiseFeedForward<B>,
+
+	/// TODO: wtf?
+	norm_a: LayerNorm<B>,
+	norm_b: LayerNorm<B>,
+
+	drop_shortcut: Dropout,
+}
+
+impl<B: Backend> TransformerBlock<B> {
+	pub fn new(cfg: &Config, device: &B::Device) -> Self {
+		Self {
+			attention: MultiheadAttention::new(
+				cfg.embed_dim,
+				cfg.head_dim,
+				cfg.n_heads,
+				cfg.context_size,
+				cfg.attention_drop_rate,
+				device,
+			),
+
+			ff: PositionWiseFeedForwardConfig::new(cfg.embed_dim, 4 * cfg.embed_dim)
+				.with_dropout(0.0)
+				.init(device),
+
+			norm_a: LayerNormConfig::new(cfg.embed_dim).init(device),
+			norm_b: LayerNormConfig::new(cfg.embed_dim).init(device),
+
+			drop_shortcut: Dropout {
+				prob: cfg.shortcut_drop_rate,
+			},
+		}
+	}
+
+	pub fn forward(&self, input: Tensor<B, 3>) -> Tensor<B, 3> {
+		let input = {
+			let shortcut = input.clone();
+			let x = self.norm_a.forward(input);
+			let x = self.attention.forward(x);
+			let x = self.drop_shortcut.forward(x);
+			x + shortcut
+		};
+
+		let input = {
+			// TODO: wtf?
+			let shortcut = input.clone();
+			let x = self.norm_b.forward(input);
+			let x = self.ff.forward(x);
+			let x = self.drop_shortcut.forward(x);
+			x + shortcut
+		};
+
+		return input;
+	}
+}