NanoDex 8m · 199,753,728 fineweb-edu tokens · loss 3.8884

README.md

@@ -0,0 +1,68 @@
+---
+license: odc-by
+datasets:
+  - HuggingFaceFW/fineweb-edu
+language:
+  - en
+library_name: transformers
+pipeline_tag: text-generation
+tags:
+  - nanodex
+  - tiny-lm
+  - pretrained-from-scratch
+---
+
+# LowOnMind-8M
+
+A **8,060,256-parameter** decoder-only language model pre-trained
+**from scratch** on [fineweb-edu](https://huggingface.co/datasets/HuggingFaceFW/fineweb-edu),
+using the [NanoDex Trainer](https://huggingface.co/spaces/hugging-science/nanodex-trainer) Space.
+
+## Architecture
+
+A standard `LlamaForCausalLM` decoder-only transformer — SiLU MLP, RMSNorm,
+rotary position embeddings, grouped-query attention, tied embeddings, no biases —
+scaled down in width and depth to fit the parameter budget.
+
+| | |
+|---|---|
+| Parameters | 8,060,256 |
+| Hidden size | 288 |
+| Layers | 9 |
+| Attention heads | 9 (KV: 3) |
+| FFN size | 704 |
+| Context length | 512 |
+| Vocab | 2,048 (custom BPE trained on fineweb-edu) |
+
+## Training
+
+| | |
+|---|---|
+| Tokens seen | 199,753,728 |
+| Steps | 381 |
+| Tokens / step | 524,288 |
+| Optimizer | AdamW(0.9, 0.95) wd=0.1 clip=1.0 |
+| LR schedule | warmup 2% + cosine to 10% (peak 1e-03) |
+| Final loss | 3.8884 (ppl 48.8) |
+| Wall time | 29.3 min |
+| Trained by | [@DedeProGames](https://huggingface.co/DedeProGames) |
+
+## Usage
+
+```python
+from transformers import AutoModelForCausalLM, AutoTokenizer
+
+tok = AutoTokenizer.from_pretrained("DedeProGames/LowOnMind-8M")
+model = AutoModelForCausalLM.from_pretrained("DedeProGames/LowOnMind-8M")
+
+ids = tok("The mitochondria is", return_tensors="pt").input_ids
+print(tok.decode(model.generate(ids, max_new_tokens=60, do_sample=True,
+                                temperature=0.8, top_k=50)[0]))
+```
+
+## Caveats
+
+This is a **nano-scale research artifact**. At this parameter count and token
+budget the model learns word shapes, common collocations and a little syntax —
+it is not a useful assistant and its output is not factual. It exists to make
+"pre-train a transformer from scratch" something you can actually watch happen.

config.json

@@ -0,0 +1,32 @@
+{
+  "architectures": [
+    "LlamaForCausalLM"
+  ],
+  "attention_bias": false,
+  "attention_dropout": 0.0,
+  "bos_token_id": 0,
+  "dtype": "float32",
+  "eos_token_id": 0,
+  "head_dim": 32,
+  "hidden_act": "silu",
+  "hidden_size": 288,
+  "initializer_range": 0.041666666666666664,
+  "intermediate_size": 704,
+  "max_position_embeddings": 512,
+  "mlp_bias": false,
+  "model_type": "llama",
+  "num_attention_heads": 9,
+  "num_hidden_layers": 9,
+  "num_key_value_heads": 3,
+  "pad_token_id": 1,
+  "pretraining_tp": 1,
+  "rms_norm_eps": 1e-05,
+  "rope_parameters": {
+    "rope_theta": 10000.0,
+    "rope_type": "default"
+  },
+  "tie_word_embeddings": true,
+  "transformers_version": "5.17.0",
+  "use_cache": true,
+  "vocab_size": 2048
+}

generation_config.json

@@ -0,0 +1,6 @@
+{
+  "bos_token_id": 0,
+  "eos_token_id": 0,
+  "pad_token_id": 1,
+  "_from_model_config": true
+}

model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ba1df7576bd7a53279bf6effcb4b621fa211f36ac3be342a669e9606989c1a85
+size 32249984

special_tokens_map.json

@@ -0,0 +1,5 @@
+{
+  "bos_token": "<|endoftext|>",
+  "eos_token": "<|endoftext|>",
+  "pad_token": "<|pad|>"
+}

tokenizer_config.json

@@ -0,0 +1,27 @@
+{
+  "added_tokens_decoder": {
+    "0": {
+      "content": "<|endoftext|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "1": {
+      "content": "<|pad|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    }
+  },
+  "bos_token": "<|endoftext|>",
+  "eos_token": "<|endoftext|>",
+  "pad_token": "<|pad|>",
+  "unk_token": null,
+  "clean_up_tokenization_spaces": false,
+  "model_max_length": 512,
+  "tokenizer_class": "PreTrainedTokenizerFast"
+}

training_run.json

@@ -0,0 +1,20 @@
+{
+  "job_id": "e6414655b517",
+  "tier": "8m",
+  "label": "NanoDex-8M",
+  "n_params": 8060256,
+  "target_tokens": 200000000,
+  "tokens_seen": 199753728,
+  "steps": 381,
+  "seq_len": 512,
+  "final_loss": 3.8884217739105225,
+  "best_loss": 3.820189207792282,
+  "peak_lr": 0.00133,
+  "batch_tokens": 524288,
+  "optimizer": "AdamW(0.9, 0.95) wd=0.1 clip=1.0",
+  "schedule": "warmup 2% + cosine to 10%",
+  "dataset": "HuggingFaceFW/fineweb-edu (sample-10BT)",
+  "architecture": "LlamaForCausalLM (SiLU, RMSNorm, RoPE, GQA, tied embeddings)",
+  "wall_time_s": 1755.7315411567688,
+  "trained_by": "DedeProGames"
+}