fixes
Browse files
README.md
CHANGED
|
@@ -1,13 +1,15 @@
|
|
| 1 |
SmolLM2, a family of compact language models available in three sizes: 135M, 360M, and 1.7B parameters.
|
| 2 |
|
| 3 |
-
In this repo is WASM compiled 1.7B model suitable for (
|
|
|
|
|
|
|
| 4 |
|
| 5 |
-
**SmolLM2-1.7B**:
|
| 6 |
Demonstrates significant improvements over its predecessor, SmolLM1-1.7B, in instruction following, knowledge, reasoning, and mathematics.
|
| 7 |
Training: Trained on 11 trillion tokens using a diverse dataset combination including FineWeb-Edu, DCLM, The Stack, and new mathematics and coding datasets.
|
| 8 |
Fine-Tuning: Developed through supervised fine-tuning (SFT) and Direct Preference Optimization (DPO) using UltraFeedback.
|
| 9 |
|
| 10 |
**Capabilities:**
|
|
|
|
| 11 |
Tasks: Supports tasks such as text rewriting, summarization, and function calling.
|
| 12 |
Datasets: Utilizes datasets developed by Argilla, such as Synth-APIGen-v0.1.
|
| 13 |
|
|
|
|
| 1 |
SmolLM2, a family of compact language models available in three sizes: 135M, 360M, and 1.7B parameters.
|
| 2 |
|
| 3 |
+
In this repo is WASM compiled 1.7B model suitable for [WebLLM](https://llm.mlc.ai/docs/deploy/webllm.html#webllm-runtime)
|
| 4 |
+
|
| 5 |
+
**SmolLM2-1.7B**
|
| 6 |
|
|
|
|
| 7 |
Demonstrates significant improvements over its predecessor, SmolLM1-1.7B, in instruction following, knowledge, reasoning, and mathematics.
|
| 8 |
Training: Trained on 11 trillion tokens using a diverse dataset combination including FineWeb-Edu, DCLM, The Stack, and new mathematics and coding datasets.
|
| 9 |
Fine-Tuning: Developed through supervised fine-tuning (SFT) and Direct Preference Optimization (DPO) using UltraFeedback.
|
| 10 |
|
| 11 |
**Capabilities:**
|
| 12 |
+
|
| 13 |
Tasks: Supports tasks such as text rewriting, summarization, and function calling.
|
| 14 |
Datasets: Utilizes datasets developed by Argilla, such as Synth-APIGen-v0.1.
|
| 15 |
|