Update README.md
Browse files
README.md
CHANGED
|
@@ -5,4 +5,39 @@ datasets:
|
|
| 5 |
language:
|
| 6 |
- en
|
| 7 |
pipeline_tag: text-generation
|
| 8 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 5 |
language:
|
| 6 |
- en
|
| 7 |
pipeline_tag: text-generation
|
| 8 |
+
tags:
|
| 9 |
+
- llama
|
| 10 |
+
- causal-lm
|
| 11 |
+
- pretrained
|
| 12 |
+
model-index:
|
| 13 |
+
- name: miniLLM-0.1B
|
| 14 |
+
results: []
|
| 15 |
+
---
|
| 16 |
+
|
| 17 |
+
# miniLLM-0.1B
|
| 18 |
+
|
| 19 |
+
A small (~109M parameters) causal language model pretrained from scratch on [OpenWebText](https://huggingface.co/datasets/Skylion007/openwebtext).
|
| 20 |
+
|
| 21 |
+
## Model Details
|
| 22 |
+
|
| 23 |
+
| Attribute | Value |
|
| 24 |
+
|---|---|
|
| 25 |
+
| Architecture | LlamaForCausalLM |
|
| 26 |
+
| Parameters | ~109M |
|
| 27 |
+
| Hidden Size | 768 |
|
| 28 |
+
| Attention Heads | 12 |
|
| 29 |
+
| Layers | 10 |
|
| 30 |
+
| Intermediate Size | 2048 |
|
| 31 |
+
| Max Sequence Length | 1024 |
|
| 32 |
+
| Vocabulary Size | 50257 |
|
| 33 |
+
| Tokenizer | GPT-2 (BPE) |
|
| 34 |
+
| Positional Encoding | RoPE (θ=10000) |
|
| 35 |
+
| Activation | SiLU |
|
| 36 |
+
| Tie Word Embeddings | Yes |
|
| 37 |
+
| Precision (training) | bfloat16 |
|
| 38 |
+
|
| 39 |
+
|
| 40 |
+
|
| 41 |
+
## Limitations
|
| 42 |
+
|
| 43 |
+
This is a small-scale pretrained model intended for research and educational purposes. It is **not** suitable for production use. Outputs may be incoherent, biased, or factually incorrect.
|