---
license: mit
pipeline_tag: text-generation
library_name: transformers
language: [
    'en', 'am', 'ar', 'as', 'az', 'be', 'bg', 'bn', 'br', 'bs', 'ca', 'cs', 'cy', 'da', 'de', 'el',
    'eo', 'es', 'et', 'eu', 'fa', 'ff', 'fi', 'fr', 'fy', 'ga', 'gd', 'gl', 'gn', 'gu', 'ha', 'he',
    'hi', 'hr', 'ht', 'hu', 'hy', 'id', 'ig', 'is', 'it', 'ja', 'jv', 'ka', 'kk', 'km', 'kn', 'ko',
    'ku', 'ky', 'la', 'lg', 'li', 'ln', 'lo', 'lt', 'lv', 'mg', 'mk', 'ml', 'mn', 'mr', 'ms', 'my',
    'ne', 'nl', 'no', 'ns', 'om', 'or', 'pa', 'pl', 'ps', 'pt', 'qu', 'rm', 'ro', 'ru', 'sa', 'si',
    'sc', 'sd', 'sk', 'sl', 'so', 'sq', 'sr', 'ss', 'su', 'sv', 'sw', 'ta', 'te', 'th', 'tl', 'tn',
    'tr', 'ug', 'uk', 'ur', 'uz', 'vi', 'wo', 'xh', 'yi', 'yo', 'zu',
]
datasets:
# core - base
- ontocord/fineweb-permissive-multilingual-2m
- distily/c4_multilingual_1M
- data-silence/sumnews
- xu-song/cc100-samples
- badrex/llm-emoji-dataset
- fblgit/simple-math
- Gusarich/math-expressions-1m
- neuralwork/arxiver
- christopher/rosetta-code
- nampdn-ai/tiny-codes
- JeanKaddour/minipile
# core - instruct
- NousResearch/hermes-function-calling-v1
- simplescaling/s1K-1.1
# base - instruct
- mlabonne/open-perfectblend
- allenai/tulu-3-sft-mixture
- rombodawg/Everything_Instruct_Multilingual
# base - reason
- open-r1/OpenR1-Math-220k
- open-thoughts/OpenThoughts-114k
- cognitivecomputations/dolphin-r1
- simplescaling/s1K-1.1
tags:
- chat
- core
- base
- instruct
- reason
---

# tangled-alpha-0.13-core

![logo](./misc/logo.jpg)

```bash
time python -B prepare_base_datasets.py
```

```
i=0, min_len=0, max_len=1073741824, block_size=8193, chunk_size=16386000, len(dataset)=1496631, len(dataset) * block_size=12261897783
Total number of tokens in the optimized dataset '../base-data-0-0-1073741824-8193-2000' is 12261897783

i=1, min_len=8193, max_len=16385, block_size=16385, chunk_size=16385000, len(dataset)=78802, len(dataset) * block_size=1291170770
Total number of tokens in the optimized dataset '../base-data-1-8193-16385-16385-1000' is 1291170770

i=2, min_len=16385, max_len=32769, block_size=32769, chunk_size=16384500, len(dataset)=23511, len(dataset) * block_size=770431959
Total number of tokens in the optimized dataset '../base-data-2-16385-32769-32769-500' is 770431959

i=3, min_len=32769, max_len=65537, block_size=65537, chunk_size=16384250, len(dataset)=5128, len(dataset) * block_size=336073736
Total number of tokens in the optimized dataset '../base-data-3-32769-65537-65537-250' is 336073736

i=4, min_len=65537, max_len=131073, block_size=131073, chunk_size=16384125, len(dataset)=1169, len(dataset) * block_size=153224337
Total number of tokens in the optimized dataset '../base-data-4-65537-131073-131073-125' is 153224337

46G     ../base-data-0-0-1073741824-8193-2000
4.9G    ../base-data-1-8193-16385-16385-1000
2.9G    ../base-data-2-16385-32769-32769-500
1.3G    ../base-data-3-32769-65537-65537-250
589M    ../base-data-4-65537-131073-131073-125
```

```bash
CUDA_VISIBLE_DEVICES=0 CUDA_LAUNCH_BLOCKING=0 PYTORCH_CUDA_ALLOC_CONF=expandable_segments:True litgpt pretrain --config pretrain_base_model_0.yaml
```

```
```

Backup `wandb`:

```bash
mv wandb wandb-pretrain-base-0
```

Copy config:

```bash
cp ../config-0.json ../out/pretrain-base-0/final/config.json
```

Chat with model:

```bash
CUDA_VISIBLE_DEVICES=0 CUDA_LAUNCH_BLOCKING=0 PYTORCH_CUDA_ALLOC_CONF=expandable_segments:True litgpt chat ../out/pretrain-base-0/final
```

```bash
CUDA_VISIBLE_DEVICES=0 CUDA_LAUNCH_BLOCKING=0 PYTORCH_CUDA_ALLOC_CONF=expandable_segments:True time litgpt evaluate --tasks 'leaderboard' --out_dir '../evaluate/pretrain-base-0/leaderboard/' --batch_size '4' --dtype 'bfloat16' '../out/pretrain-base-0/final'
```

```
```

```bash
litgpt convert_pretrained_checkpoint ../out/pretrain-base-0/final ../out/pretrain-base-0/checkpoint
```

```bash
CUDA_VISIBLE_DEVICES=0 CUDA_LAUNCH_BLOCKING=0 PYTORCH_CUDA_ALLOC_CONF=expandable_segments:True litgpt pretrain --config pretrain_base_model_1.yaml
```

```bash
litgpt convert_pretrained_checkpoint ../out/pretrain-base-1/final ../out/pretrain-base-1/checkpoint
```

```bash
CUDA_VISIBLE_DEVICES=0 CUDA_LAUNCH_BLOCKING=0 PYTORCH_CUDA_ALLOC_CONF=expandable_segments:True litgpt pretrain --config pretrain_base_model_2.yaml
```

```bash
litgpt convert_pretrained_checkpoint ../out/pretrain-base-2/final ../out/pretrain-base-2/checkpoint
```

```bash
CUDA_VISIBLE_DEVICES=0 CUDA_LAUNCH_BLOCKING=0 PYTORCH_CUDA_ALLOC_CONF=expandable_segments:True litgpt pretrain --config pretrain_base_model_3.yaml
```

```bash
CUDA_VISIBLE_DEVICES=0 CUDA_LAUNCH_BLOCKING=0 PYTORCH_CUDA_ALLOC_CONF=expandable_segments:True time litgpt evaluate --tasks 'leaderboard' --out_dir '../evaluate/pretrain-base-3/leaderboard/' --batch_size '4' --dtype 'bfloat16' '../out/pretrain-base-3/final'
```

```
```