We lifted 9 functions out of this paper's own repositories and ran 9 of them in a sandbox. "Ran" means the function executed on a synthesized input and returned a value. It is not a reproduction of the paper's results.
| Repository | Role | Ran |
|---|---|---|
| Pints-AI/1.5-Pints | canonical | 9 of 9 |
| Function | Status | Where it lives |
|---|---|---|
| apply_rope | Ran | Pints-AI/1.5-Pints/lit_gpt/model.py code served (permissive licence) · get_code("5728c74084e12d35") |
| build_rope_cache | Ran | Pints-AI/1.5-Pints/lit_gpt/model.py code served (permissive licence) · get_code("a30c7463cb6330d7") |
| extend_checkpoint_dir | Ran | Pints-AI/1.5-Pints/convert/convert_hf_to_lit.py code served (permissive licence) · get_code("c4561ab3f1e4d51b") |
| generate | Ran | Pints-AI/1.5-Pints/inference/generate.py code served (permissive licence) · get_code("aee31819bbec4f25") |
| get_lr_scheduler | Ran | Pints-AI/1.5-Pints/finetune/full.py code served (permissive licence) · get_code("92caa4cd63a752cc") |
| layer_template | Ran | Pints-AI/1.5-Pints/convert/convert_hf_to_lit.py code served (permissive licence) · get_code("a41f9d163fa603c7") |
| layer_template | Ran | Pints-AI/1.5-Pints/convert/convert_lit_to_hf.py code served (permissive licence) · get_code("4c9b2b019395cbbd") |
| rename | Ran | Pints-AI/1.5-Pints/convert/bin_to_safetensors.py code served (permissive licence) · get_code("f732d3a613a637b6") |
| shared_pointers | Ran | Pints-AI/1.5-Pints/convert/bin_to_safetensors.py code served (permissive licence) · get_code("5bfc51991538ca34") |
Some links come from the archived Papers with Code dataset (CC BY-SA 4.0): attribution and licence.
This paper presents a compute-efficient approach to pre-training a Language Model-the "1.5-Pints"-in only 9 days, while outperforming state-of-the-art models as an instruction-following assistant.Based on MT-Bench (a benchmark that emulates human judgments), 1.5-Pints outperforms Apple's OpenELM and Microsoft's Phi.This is achieved by a carefully curated pre-training dataset of 57 billion tokens, using a mix of automated workflows and manual human review. The selection of the dataset prioritizes content that is considered expository and "textbook-like" to aid the model in reasoning and logical deduction, culminating in its overall ability as a strong and versatile AI model. In terms of the model architecture, we employed a modified Mistral tokenizer, alongside a Llama-2 architecture for wider compatibility. For training, we adopted the methodologies used by StableLM, TinyLlama, and Huggingface Zephyr. 1.5-Pints demonstrates that by focusing on data quality over quantity in LLM training, we can significantly reduce training time and resources required. We believe this approach will not only make pre-training more accessible but also reduce our carbon footprint. Our findings and resources from this research are open-sourced, aiming to facilitate further advancements in the field. The 1.5-Pints model is available in two versions: 2K and 16K context windows.
The same record, over MCP at https://syntology.ai/mcp:
get_harvested_code_for_paper("2408.03506")
get_code_for_paper("2408.03506")
have("2408.03506")
Connect an agent — have() is free.