- Updated MiniMind-V1 (108M) model, using minimind_tokenizer, with 3 rounds of pre-training + 10 rounds of SFT, more thorough training, stronger performance. - 更新MiniMind-V1 (108M)模型,采用minimind_tokenizer,预训练轮次3 + SFT轮次10,更充分训练,性能更强。 - The project has been depl...
- 更新MiniMind-V1 (108M)模型,采用minimind_tokenizer,预训练轮次3 + SFT轮次10,更充分训练,性能更强。 2024-08-27 - 项目首次开源 @@ -115,30 +119,30 @@ python 2-eval.py * 2.6 `python 4-lora_sft.py` 执行lora微调(非必须)。 * 2.7 `python 5-dpo_train.py` 执行DPO人类偏好强化...
minimind-v1-small / data_process.py data_process.py 5.97 KB 一键复制 编辑 原始数据 按行查看 历史 风始于青萍末 提交于 2个月前 . Initial commit 12345678910111213141516171819202122232425262728293031323334353637383940414243444546474849505152535455565758596061626364656667...
print(f'模型参数:{count_parameters(lm_model)/1e6}百万 ={count_parameters(lm_model)/1e9}B (Billion)') # 保存模型 lm_model.save_pretrained("minimind-v1-small",safe_serialization=False) # 定义一个函数,用于导出tokenizer defexport_tokenizer(): ...
#INTO1尹浩宇欧美蜜嗓绝美转音# ️#INTO1尹浩宇语言天才# 尹浩宇,喜欢你真的好幸福。你的一举一动,你的善良,时刻让我提醒自己要做和你一样优秀的人。我希望你可以开心健康快乐,拥有你想要的那种人生,做自己想做的,在派大星满满的爱里,永远星途璀璨 @INTO1-尹浩宇 ...
minimind-v1 jingyaogong committed Sep 1, 2024 1 parent c8f957d commit 7e9621f Showing 1 changed file with 128 additions and 0 deletions. Whitespace Ignore whitespace Split Unified 128 changes: 128 additions & 0 deletions 128 fast_infenence.py Original file line numberDiff line numberDiff...
minimind-v1 jingyaogong committed Sep 1, 2024 1 parent 0047679 commit 6d1e7ca Showing 2 changed files with 12 additions and 0 deletions. Whitespace Ignore whitespace Split Unified README.md README_en.md 7 changes: 7 additions & 0 deletions 7 README.md Original file line numberDiff line...