Duplicate from Mxode/Chinese-Instruct

Co-authored-by: Max Zhang <Mxode@users.noreply.huggingface.co>

.gitattributes

@@ -0,0 +1,75 @@
+*.7z filter=lfs diff=lfs merge=lfs -text
+*.arrow filter=lfs diff=lfs merge=lfs -text
+*.bin filter=lfs diff=lfs merge=lfs -text
+*.bz2 filter=lfs diff=lfs merge=lfs -text
+*.ckpt filter=lfs diff=lfs merge=lfs -text
+*.ftz filter=lfs diff=lfs merge=lfs -text
+*.gz filter=lfs diff=lfs merge=lfs -text
+*.h5 filter=lfs diff=lfs merge=lfs -text
+*.joblib filter=lfs diff=lfs merge=lfs -text
+*.lfs.* filter=lfs diff=lfs merge=lfs -text
+*.lz4 filter=lfs diff=lfs merge=lfs -text
+*.mds filter=lfs diff=lfs merge=lfs -text
+*.mlmodel filter=lfs diff=lfs merge=lfs -text
+*.model filter=lfs diff=lfs merge=lfs -text
+*.msgpack filter=lfs diff=lfs merge=lfs -text
+*.npy filter=lfs diff=lfs merge=lfs -text
+*.npz filter=lfs diff=lfs merge=lfs -text
+*.onnx filter=lfs diff=lfs merge=lfs -text
+*.ot filter=lfs diff=lfs merge=lfs -text
+*.parquet filter=lfs diff=lfs merge=lfs -text
+*.pb filter=lfs diff=lfs merge=lfs -text
+*.pickle filter=lfs diff=lfs merge=lfs -text
+*.pkl filter=lfs diff=lfs merge=lfs -text
+*.pt filter=lfs diff=lfs merge=lfs -text
+*.pth filter=lfs diff=lfs merge=lfs -text
+*.rar filter=lfs diff=lfs merge=lfs -text
+*.safetensors filter=lfs diff=lfs merge=lfs -text
+saved_model/**/* filter=lfs diff=lfs merge=lfs -text
+*.tar.* filter=lfs diff=lfs merge=lfs -text
+*.tar filter=lfs diff=lfs merge=lfs -text
+*.tflite filter=lfs diff=lfs merge=lfs -text
+*.tgz filter=lfs diff=lfs merge=lfs -text
+*.wasm filter=lfs diff=lfs merge=lfs -text
+*.xz filter=lfs diff=lfs merge=lfs -text
+*.zip filter=lfs diff=lfs merge=lfs -text
+*.zst filter=lfs diff=lfs merge=lfs -text
+*tfevents* filter=lfs diff=lfs merge=lfs -text
+# Audio files - uncompressed
+*.pcm filter=lfs diff=lfs merge=lfs -text
+*.sam filter=lfs diff=lfs merge=lfs -text
+*.raw filter=lfs diff=lfs merge=lfs -text
+# Audio files - compressed
+*.aac filter=lfs diff=lfs merge=lfs -text
+*.flac filter=lfs diff=lfs merge=lfs -text
+*.mp3 filter=lfs diff=lfs merge=lfs -text
+*.ogg filter=lfs diff=lfs merge=lfs -text
+*.wav filter=lfs diff=lfs merge=lfs -text
+# Image files - uncompressed
+*.bmp filter=lfs diff=lfs merge=lfs -text
+*.gif filter=lfs diff=lfs merge=lfs -text
+*.png filter=lfs diff=lfs merge=lfs -text
+*.tiff filter=lfs diff=lfs merge=lfs -text
+# Image files - compressed
+*.jpg filter=lfs diff=lfs merge=lfs -text
+*.jpeg filter=lfs diff=lfs merge=lfs -text
+*.webp filter=lfs diff=lfs merge=lfs -text
+# Video files - compressed
+*.mp4 filter=lfs diff=lfs merge=lfs -text
+*.webm filter=lfs diff=lfs merge=lfs -text
+stem_zh/stem_zh.jsonl filter=lfs diff=lfs merge=lfs -text
+infinity-instruct/train.jsonl filter=lfs diff=lfs merge=lfs -text
+dpsk-r1-distil/train.jsonl filter=lfs diff=lfs merge=lfs -text
+firefly/train.jsonl filter=lfs diff=lfs merge=lfs -text
+magpie/train.jsonl filter=lfs diff=lfs merge=lfs -text
+coig-cqia/train.jsonl filter=lfs diff=lfs merge=lfs -text
+disc-law/train.jsonl filter=lfs diff=lfs merge=lfs -text
+neo_sft_phase2/train.jsonl filter=lfs diff=lfs merge=lfs -text
+chinese-medical/medical-train.jsonl filter=lfs diff=lfs merge=lfs -text
+chinese-reasoning-distil/train-1.jsonl filter=lfs diff=lfs merge=lfs -text
+psycho-10k-dpsk-r1/train.jsonl filter=lfs diff=lfs merge=lfs -text
+sof-c-zh/train.jsonl filter=lfs diff=lfs merge=lfs -text
+chinese-reasoning-distil/chinese-reasoning-distil/train-2.jsonl filter=lfs diff=lfs merge=lfs -text
+chinese-reasoning-distil/train-2.jsonl filter=lfs diff=lfs merge=lfs -text
+industryinstruction/train.jsonl filter=lfs diff=lfs merge=lfs -text
+Chinese-QA-AFAF/Chinese-QA-AFAF.jsonl filter=lfs diff=lfs merge=lfs -text

Chinese-QA-AFAF/Chinese-QA-AFAF.jsonl

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:98275b5d470f2ee1b4ccf795b9c4a41dfa6ecedf75c8c4c656c999dc7d8c2b78
+size 1263196430

README.md

@@ -0,0 +1,141 @@
+---
+configs:
+- config_name: stem_zh
+  data_files:
+  - split: train
+    path: stem_zh/**/*
+- config_name: infinity-instruct
+  data_files:
+  - split: train
+    path: infinity-instruct/**/*
+- config_name: firefly
+  data_files:
+  - split: train
+    path: firefly/**/*
+- config_name: magpie
+  data_files:
+  - split: train
+    path: magpie/**/*
+- config_name: dpsk-r1-distil
+  default: true
+  data_files:
+  - split: train
+    path: dpsk-r1-distil/**/*
+- config_name: coig-cqia
+  data_files:
+  - split: train
+    path: coig-cqia/**/*
+- config_name: disc-law
+  data_files:
+  - split: train
+    path: disc-law/**/*
+- config_name: neo_sft_phase2
+  data_files:
+  - split: train
+    path: neo_sft_phase2/**/*
+- config_name: chinese-medical
+  data_files:
+  - split: train
+    path: chinese-medical/**/*
+- config_name: chinese-reasoning-distil
+  data_files:
+  - split: train
+    path: chinese-reasoning-distil/**/*
+- config_name: psycho-10k-dpsk-r1
+  data_files:
+  - split: train
+    path: psycho-10k-dpsk-r1/**/*
+- config_name: sof-c-zh
+  data_files:
+  - split: train
+    path: sof-c-zh/**/*
+- config_name: industryinstruction
+  data_files:
+  - split: train
+    path: industryinstruction/**/*
+- config_name: Chinese-QA-AFAF
+  data_files:
+  - split: train
+    path: Chinese-QA-AFAF/**/*
+license: cc-by-sa-4.0
+task_categories:
+- text-generation
+- question-answering
+language:
+- zh
+---
+
+<h1 align="center">
+  中文指令微调数据集
+</h1>
+
+<p align="center">
+  <a href="https://github.com/Mxoder/Maxs-Awesome-Datasets" target="_blank">💻 Github Repo</a> <br>
+</p>
+
+本项目旨在构建一个高质量、多领域、大规模的中文指令微调数据集。
+
+本项目将会**持续更新**。更多数据集欢迎访问 [Github Repo](https://github.com/Mxoder/Maxs-Awesome-Datasets)。
+
+> [!TIP]
+> 如果您想要一个可用于学习的**简化版**中文指令数据集，可以访问：[Mxode/Chinese-Instruct-Lite](https://huggingface.co/datasets/Mxode/Chinese-Instruct-Lite)
+
+---
+
+## 具体构成
+
+- `dpsk-r1-distil`：中文 DeepSeek-R1 蒸馏数据集，来自 [Congliu/Chinese-DeepSeek-R1-Distill-data-110k](https://huggingface.co/datasets/Congliu/Chinese-DeepSeek-R1-Distill-data-110k)，根据打分质量做了筛选，提取了最终的回答，未包含思考过程。
+
+- `chinese-reasoning-distil`：中文推理蒸馏数据集，来自 [Mxode/Chinese-Reasoning-Distil-Data](https://huggingface.co/datasets/Mxode/Chinese-Reasoning-Distil-Data)，提取了最终的回答，未包含思考过程。
+
+- `firefly`：中文通用指令微调数据集，指令取自 [Mxode/Firefly-1.1M-Rephrased](https://huggingface.co/datasets/Mxode/Firefly-1.1M-Rephrased)，其本身已经相较于原 Firefly 数据集做了改进优化，此处再次针对所有优化的指令用更强的模型（如 DeepSeek-V2.5）做了新的回答。
+
+- `stem_zh`：中文 STEM 指令微调数据集，指令取自 [hfl/stem_zh_instruction](https://huggingface.co/datasets/hfl/stem_zh_instruction)，以原数据集的指令和答案重新优化合成了新的回复数据，并去除了一些幻觉数据。
+
+- `magpie`：中文通用指令微调数据集，从 [Magpie](https://huggingface.co/Magpie-Align) 项目采样筛选指令，再利用 GPT-4o 进行指令改写和筛选，最后再用 GPT-4o-mini 生成回复，数目有 500K+。
+
+- `infinity-instruct`：中文通用指令微调数据集，从原项目 [BAAI/Infinity-Instruct](https://huggingface.co/datasets/BAAI/Infinity-Instruct) 筛选出了所有的中文数据（重新做了语言检测，其自带的语言元数据仍有误），然后提取了其中所有的单轮对话。
+
+- `neo_sft_phase2`：中文通用指令微调数据集，从原项目 [m-a-p/neo_sft_phase2](https://huggingface.co/datasets/m-a-p/neo_sft_phase2) 检测筛选了所有的中文单轮数据，并做了启发式过滤。
+
+- `coig-cqia`：中文通用指令微调数据集，从原项目 [m-a-p/COIG-CQIA](https://huggingface.co/datasets/m-a-p/COIG-CQIA) 筛选了所有经过人工校验的数据，并做了启发式过滤。
+
+- `disc-law`：中文法律指令微调数据集，选取了 [ShengbinYue/DISC-Law-SFT](https://huggingface.co/datasets/ShengbinYue/DISC-Law-SFT) 其中的 `DISC-Law-SFT-Pair` 数据，并做了启发式过滤。
+
+- `chinese-medical`：中文医疗指令微调数据集，来自项目 [Mxode/Chinese-Medical-Instruct-1M](https://huggingface.co/datasets/Mxode/Chinese-Medical-Instruct-1M)。
+
+- `psycho-10k-dpsk-r1`：中文心理学 DeepSeek-R1 蒸馏数据集，来自 [Kedreamix/psychology-10k-Deepseek-R1-zh](https://huggingface.co/datasets/Kedreamix/psychology-10k-Deepseek-R1-zh)，其应是 [samhog/psychology-10k](https://huggingface.co/datasets/samhog/psychology-10k) 数据集的翻译再合成，本数据集提取了最终的回答，未包含思考过程。
+
+- `sof-c-zh`：中文代码指令微调数据集，来自 [Mxode/Chinese-StackOverflow-QA-C_Language](https://huggingface.co/datasets/Mxode/Chinese-StackOverflow-QA-C_Language)，数据源为 StackOverflow 的 C 语言问答。
+
+- `industryinstruction`：中文行业指令数据集，来自 [BAAI/IndustryInstruction](https://huggingface.co/datasets/BAAI/IndustryInstruction)，对原数据做了清洗，具体见 [Mxode/IndustryInstruction-Chinese](https://huggingface.co/datasets/Mxode/IndustryInstruction-Chinese)，此处选取了 DEITA 分数 / reward 分数在前 50% 的数据。
+
+- `Chinese-QA-AFAF`：中文农林牧渔问答数据集，来自 [Mxode/Chinese-QA-Agriculture_Forestry_Animal_Husbandry_Fishery](https://huggingface.co/datasets/Mxode/Chinese-QA-Agriculture_Forestry_Animal_Husbandry_Fishery)，在原数据的基础上做了二次过滤和清洗。
+
+---
+
+## 如何使用
+
+目前数据集均整理为了 `prompt`-`response` 的问答对形式，所有子集格式统一。
+
+可以指定加载其中的一个子集：
+
+```python
+from datasets import load_dataset
+
+# 加载其中的一个子集
+subset = "dpsk-r1-distil"
+ds = load_dataset("Mxode/Chinese-Instruct", subset)
+```
+
+也可以加载多个子集或者全部子集，可以通过如下方式获取子集名称：
+
+```python
+from datasets import get_dataset_config_names
+
+configs = get_dataset_config_names("Mxode/Chinese-Instruct")
+
+print(configs)
+```
+
+这样会列出所有子集的名称，可以按需加载。

chinese-medical/medical-train.jsonl

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:eadb8aa9d5424dce6dc0d6fecfcb056d9454d1a890c6ef28a9f1f056b6b28746
+size 1685448521

chinese-reasoning-distil/train-1.jsonl

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:23e511328ae82e43bf14a3c4416e43081bb4bd987f08a97c083c778610e57164
+size 241406083

chinese-reasoning-distil/train-2.jsonl

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:02eeaea9fd684fff6e006b56a450eddadfc37cce9a18861007193c959c418503
+size 414110512

coig-cqia/train.jsonl

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e79bbdea5e4711588f0ab9c97ca25a97a85a0640904e080b7e1ac59ac2eebf19
+size 20321394

disc-law/train.jsonl

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f4260e812e2c0910fac6d52f03f7bd30d8640053ce76e90610ea2b1ad8f46f63
+size 151249787

dpsk-r1-distil/train.jsonl

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b8220841c44b77e3976345bb2a8af0affac02dc215e56c2fdaa29244475ecd42
+size 220783740

firefly/train.jsonl

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:aedf54a03ccddc9545ef98627d10e76fc7940c3aaaf555186e5d7fd1f754f7a4
+size 2284291145

industryinstruction/train.jsonl

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:cce11064dadf30140d08e53ec1113b6884543bfc4219efc8a8c3b50fc1105abc
+size 331079592

infinity-instruct/train.jsonl

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:599c60efa797baa1cf7e0e3026151ac63db76c0a06e5f641a59517fd876bb5c5
+size 540802587

magpie/train.jsonl

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:780018d3dd85141772068156aad7cb113280d8655b902c70dc319e090e78ce31
+size 1381853459

neo_sft_phase2/train.jsonl

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:6ecd1ee0ebf8697f1f2c5968edad8b85fae1a6faa7bb54dfd4212b923e0c66ac
+size 91798265

psycho-10k-dpsk-r1/train.jsonl

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:44b48de68589703d62059a9072ccd56f63bf5d2681c4e6ff911de984351dbf60
+size 26636968

sof-c-zh/train.jsonl

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9e9ca41bf96687afa35f8beb8770706f8166b5c9d143412281a8f10d8c32b224
+size 451323523

stem_zh/stem_zh.jsonl

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:036962df37381032c6fd183d0407e9216618c276fab1a1aa9b33c6a47b03aa24
+size 421690670