File size: 7,707 Bytes
aa3d3f0 fbd85d7 aa3d3f0 fbd85d7 aa3d3f0 fbd85d7 aa3d3f0 fbd85d7 aa3d3f0 fbd85d7 aa3d3f0 fbd85d7 aa3d3f0 fbd85d7 aa3d3f0 fbd85d7 aa3d3f0 fbd85d7 aa3d3f0 fbd85d7 aa3d3f0 fbd85d7 aa3d3f0 fbd85d7 aa3d3f0 fbd85d7 aa3d3f0 fbd85d7 aa3d3f0 fbd85d7 aa3d3f0 fbd85d7 aa3d3f0 fbd85d7 aa3d3f0 fbd85d7 aa3d3f0 | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 | ---
license: mit
language:
- en
tags:
- julia
- lux
- transformer
- monarch-mixer
- language-model
- chinchilla
- bpe
datasets:
- LisaMegaWatts/philosophy-corpus
pipeline_tag: text-generation
---
# Julia SLM β Small Language Models in Pure Julia
Transformer and Monarch Mixer language models built entirely in Julia using [Lux.jl](https://github.com/LuxDL/Lux.jl), trained on the [philosophy-corpus](https://huggingface.co/datasets/LisaMegaWatts/philosophy-corpus) dataset.
## Models
### Head-to-Head Comparison
| Metric | Transformer (`5m-chinchilla/`) | Monarch Mixer (`5m-monarch/`) |
|--------|------|------|
| Parameters | 5,037,312 (5.04M) | 4,983,040 (4.98M) |
| Blocks | 6 | 8 |
| Sequence mixing | Softmax attention (4 heads) | Multi-head Monarch (8 heads) + causal conv |
| Channel mixing | SwiGLU (256β640β256) | SwiGLU (256β640β256) |
| Positional encoding | RoPE | None (learned via Monarch factors) |
| **Val loss** | **3.54** | **3.65** |
| **Val PPL** | **34.5** | **38.4** |
| Training time | 66 min | 89 min |
| Throughput | ~26K tok/s | ~19K tok/s |
Both trained identically: AdamW (lr=6e-4), cosine decay, 12,305 steps, batch 32, RTX 3060 12GB.
---
### 5M Chinchilla Transformer (`5m-chinchilla/`)
5.04M parameter decoder-only transformer trained to Chinchilla-optimal (100M tokens at 20 tokens/param).
| Param | Value |
|-------|-------|
| Parameters | 5,037,312 |
| Architecture | Decoder-only Transformer |
| Embedding dim | 256 |
| Layers | 6 |
| Attention heads | 4 |
| Head dim | 64 |
| FFN multiplier | 4x (SwiGLU) |
| Context length | 256 |
| Vocab size | 2,000 (BPE) |
| Weight tying | Yes |
| Normalization | RMSNorm (pre-norm) |
| Positional encoding | RoPE |
**Loss curve:**
| Step | Train Loss | Val Loss | Val PPL |
|------|-----------|----------|---------|
| 500 | 6.69 | 5.01 | 149.6 |
| 2,000 | 4.09 | 4.02 | 56.0 |
| 6,000 | 3.72 | 3.70 | 40.4 |
| 10,000 | 3.58 | 3.57 | 35.4 |
| 12,305 | 3.55 | 3.54 | 34.5 |
---
### 5M Monarch Mixer (`5m-monarch/`)
4.98M parameter Monarch Mixer variant using sub-quadratic sequence mixing with structured matrices.
| Param | Value |
|-------|-------|
| Parameters | 4,983,040 |
| Architecture | Monarch Mixer |
| Embedding dim | 256 |
| Layers | 8 |
| Monarch heads | 8 |
| Conv kernel | 4 (causal depthwise) |
| FFN multiplier | 4x (SwiGLU) |
| Context length | 256 |
| Vocab size | 2,000 (BPE) |
| Weight tying | Yes |
| Normalization | RMSNorm (pre-norm) |
| Gating | Learned sigmoid gate |
**How Monarch Mixer works:**
A Monarch matrix of size TΓT (T=pΒ²=256, p=16) factorizes as:
```
M = Pα΅ Β· BlockDiag(L1) Β· P Β· BlockDiag(L2)
```
where L1, L2 are p block-diagonal matrices of size pΓp, and P is a reshape-transpose permutation. Parameters: 2pΒ³ = 2T^{3/2} (8,192 vs 65,536 for dense).
Each block uses 8 independent Monarch heads (each mixing 32 channels over 256 positions) combined with a causal depthwise convolution for local n-gram patterns, gated by a learned sigmoid.
**Loss curve:**
| Step | Train Loss | Val Loss | Val PPL |
|------|-----------|----------|---------|
| 500 | 6.31 | 5.26 | 192.4 |
| 2,000 | 4.15 | 4.15 | 63.4 |
| 6,000 | 3.77 | 3.79 | 44.3 |
| 10,000 | 3.62 | 3.67 | 39.3 |
| 12,305 | 3.62 | 3.65 | 38.4 |
**Key findings:**
- Monarch reaches **94% of baseline quality** (3.65 vs 3.54 val loss) with O(T^{3/2}) parameter complexity in sequence mixing
- Uses **4x fewer parameters per block** in sequence mixing (67K vs 262K), enabling 8 blocks instead of 6
- Generates coherent English text with dialogue, grammar, and narrative structure
- First known Julia implementation of Monarch Mixer for language modeling
## Architecture
### Transformer
```
JuliaGPTModel
βββ tok_emb: Embedding(2000 β 256) # weight-tied with output head
βββ rope: RotaryPositionalEncoding(256)
βββ blocks Γ 6:
β βββ ln1: RMSNorm(256)
β βββ attn: MultiHeadAttention(4 heads, 64 dim each)
β β βββ wq, wk, wv: Dense(256 β 256)
β β βββ wo: Dense(256 β 256)
β βββ ln2: RMSNorm(256)
β βββ ffn: SwiGLU(256 β 640 β 256)
βββ ln_f: RMSNorm(256)
βββ head: TiedEmbeddingHead β (2000,)
```
### Monarch Mixer
```
JuliaGPTModel
βββ tok_emb: Embedding(2000 β 256) # weight-tied with output head
βββ blocks Γ 8:
β βββ ln1: RMSNorm(256)
β βββ seq_mixer: MonarchSequenceMixer
β β βββ conv: CausalDepthwiseConv1d(256, kernel=4)
β β βββ monarchs Γ 8: MonarchMatrix(256, L1/L2 β β^{16Γ16Γ16})
β β βββ gate: LearnedGate(256)
β βββ ln2: RMSNorm(256)
β βββ ffn: SwiGLU(256 β 640 β 256)
βββ ln_f: RMSNorm(256)
βββ head: TiedEmbeddingHead β (2000,)
```
## Usage
### Load and generate (Transformer)
```julia
using Pkg; Pkg.activate("julia-slm")
include("src/JuliaGPT.jl")
using .JuliaGPT
using .JuliaGPT: Lux, CUDA
tok = BPETokenizer("path/to/vocab.json", "path/to/merges.txt")
device = Lux.gpu_device()
ps, st, _, step, val_loss = load_checkpoint("5m-chinchilla/final.jld2"; device)
model = create_model(ModelConfig(;
vocab_size=vocab_size(tok), embed_dim=256, n_layers=6,
n_heads=4, head_dim=64, ffn_mult=4, context_length=256,
weight_tying=true,
))
text = generate(model, ps, st, tok, "the nature of ";
max_new_tokens=200, temperature=0.8, top_k=40)
println(text)
```
### Load and generate (Monarch Mixer)
```julia
ps, st, _, step, val_loss = load_checkpoint("5m-monarch/final.jld2"; device)
model = create_model(ModelConfig(;
arch="monarch",
vocab_size=vocab_size(tok), embed_dim=256, n_layers=8,
n_heads=4, head_dim=64, ffn_mult=4, context_length=256,
weight_tying=true, n_monarch_heads=8, conv_kernel_size=4,
))
text = generate(model, ps, st, tok, "the nature of ";
max_new_tokens=200, temperature=0.8, top_k=40)
println(text)
```
### Train from scratch
```bash
# Transformer baseline
julia --project scripts/train.jl --config config/5m.toml
# Monarch Mixer
julia --project scripts/train.jl --config config/5m-monarch.toml
```
## Dataset
Trained on [LisaMegaWatts/philosophy-corpus](https://huggingface.co/datasets/LisaMegaWatts/philosophy-corpus) β 981 source texts (BookCorpus, WikiText-103, PG-19, classical philosophy) processed through a custom text pipeline with deduplication and quality scoring.
- **Train tokens**: 794.9M (pre-encoded as `train.bin`)
- **Val tokens**: 88.2M (pre-encoded as `val.bin`)
- **Tokenizer**: ByteLevel BPE, 2,000 vocab
## Framework
Built with:
- [Lux.jl](https://github.com/LuxDL/Lux.jl) β Explicit-parameter neural networks
- [Zygote.jl](https://github.com/FluxML/Zygote.jl) β Automatic differentiation
- [CUDA.jl](https://github.com/JuliaGPU/CUDA.jl) β GPU acceleration
- [NNlib.jl](https://github.com/FluxML/NNlib.jl) β Batched matrix multiply, activations
- [Optimisers.jl](https://github.com/FluxML/Optimisers.jl) β AdamW with cosine LR
## Files
```
5m-chinchilla/ # Baseline transformer
βββ config.toml
βββ final.jld2 # Step 12,305
βββ step_12000.jld2
5m-monarch/ # Monarch Mixer variant
βββ config.toml
βββ final.jld2 # Step 12,305
βββ step_12000.jld2
```
Checkpoints are JLD2 format containing: model parameters (`ps`), model state (`st`), optimizer state, step number, and best validation loss.
## References
- [Monarch Mixer (Dao et al., 2023)](https://arxiv.org/abs/2310.12109) β Sub-quadratic GEMM-based architecture
- [Chinchilla (Hoffmann et al., 2022)](https://arxiv.org/abs/2203.15556) β Compute-optimal training scaling
## License
MIT
|