apunkt/PageIndex
1
0
Fork 0
mirror of https://github.com/VectifyAI/PageIndex.git synced 2026-04-24 23:56:21 +02:00
Code Issues Projects Releases Packages Wiki Activity Actions 2

fix notebook format

Browse source
This commit is contained in:
Ray 2025-08-22 01:11:48 +08:00
parent e0edea6d51
commit cf0a599cff
1 changed files with 37 additions and 25 deletions
Download patch file Download diff file Expand all files Collapse all files

62
cookbook/pageindex_RAG_simple.ipynb
View file

@ -124,7 +124,7 @@
},
{
"cell_type": "code",
"execution_count": 19,
"execution_count": 51,
"metadata": {
"id": "hmj3POkDcG1N"
},
@ -154,7 +154,8 @@
" pprint(cleaned_tree, sort_dicts=False, width=100)\n",
"\n",
"def show(text, width=100):\n",
" print(textwrap.fill(text, width=width))\n",
" for line in text.splitlines():\n",
" print(textwrap.fill(line, width=width))\n",
"\n",
"def create_node_mapping(tree):\n",
" \"\"\"Create a mapping of node_id to node for quick lookup\"\"\"\n",
@ -233,7 +234,7 @@
},
{
"cell_type": "code",
"execution_count": 20,
"execution_count": 61,
"metadata": {
"colab": {
"base_uri": "https://localhost:8080/",
@ -390,7 +391,7 @@
},
{
"cell_type": "code",
"execution_count": 35,
"execution_count": 57,
"metadata": {
"colab": {
"base_uri": "https://localhost:8080/",
@ -449,7 +450,7 @@
},
{
"cell_type": "code",
"execution_count": 37,
"execution_count": 58,
"metadata": {
"colab": {
"base_uri": "https://localhost:8080/",
@ -464,17 +465,23 @@
"output_type": "stream",
"text": [
"Retrieved Context:\n",
"## 5. Conclusion, Limitations, and Future Work In this work, we share our journey in enhancing\n",
"model reasoning abilities through reinforcement learning. DeepSeek-R1-Zero represents a pure RL\n",
"approach without relying on cold-start data, achieving strong performance across various tasks.\n",
"DeepSeek-R1 is more powerful, leveraging cold-start data alongside iterative RL fine-tuning.\n",
"Ultimately, DeepSeek-R1 achieves performance comparable to OpenAI-o1-1217 on a range of tasks. We\n",
"further explore distillation the reasoning capability to small dense models. We use DeepSeek-R1 as\n",
"the teacher model to generate 800K training samples, and fine-tune several small dense models. The\n",
"results are promising: DeepSeek-R1-Distill-Qwen-1.5B outperforms GPT-4o and Claude-3.5-Sonnet on\n",
"\n",
"## 5. Conclusion, Limitations, and Future Work\n",
"\n",
"In this work, we share our journey in enhancing model reasoning abilities through reinforcement\n",
"learning. DeepSeek-R1-Zero represents a pure RL approach without relying on cold-start data,\n",
"achieving strong performance across various tasks. DeepSeek-R1 is more powerful, leveraging cold-\n",
"start data alongside iterative RL fine-tuning. Ultimately, DeepSeek-R1 achieves performance\n",
"comparable to OpenAI-o1-1217 on a range of tasks.\n",
"\n",
"We further explore distillation the reasoning capability to small dense models. We use DeepSeek-R1\n",
"as the teacher model to generate 800K training samples, and fine-tune several small dense models.\n",
"The results are promising: DeepSeek-R1-Distill-Qwen-1.5B outperforms GPT-4o and Claude-3.5-Sonnet on\n",
"math benchmarks with $28.9 \\%$ on AIME and $83.9 \\%$ on MATH. Other dense models also achieve\n",
"impressive results, significantly outperforming other instructiontuned models based on the same\n",
"underlying checkpoints. In the fut...\n"
"underlying checkpoints.\n",
"\n",
"In the fut...\n"
]
}
],
@ -482,7 +489,7 @@
"node_list = json.loads(tree_search_result)[\"node_list\"]\n",
"relevant_content = \"\\n\\n\".join(node_map[node_id][\"text\"] for node_id in node_list)\n",
"\n",
"print('Retrieved Context:')\n",
"print('Retrieved Context:\\n')\n",
"show(relevant_content[:1000] + '...')"
]
},
@ -495,7 +502,7 @@
},
{
"cell_type": "code",
"execution_count": 36,
"execution_count": 59,
"metadata": {
"colab": {
"base_uri": "https://localhost:8080/",
@ -510,14 +517,19 @@
"output_type": "stream",
"text": [
"Generated Answer:\n",
"**Conclusions in this document:** - DeepSeek-R1-Zero, a pure reinforcement learning (RL) model\n",
"without cold-start data, achieves strong performance across various tasks. - DeepSeek-R1, which\n",
"combines cold-start data with iterative RL fine-tuning, is even more powerful and achieves\n",
"performance comparable to OpenAI-o1-1217 on a range of tasks. - Distilling DeepSeek-R1’s reasoning\n",
"capabilities into smaller dense models is effective: DeepSeek-R1-Distill-Qwen-1.5B outperforms\n",
"GPT-4o and Claude-3.5-Sonnet on math benchmarks, and other dense models also show significant\n",
"improvements over similar instruction-tuned models. - Overall, the approaches described demonstrate\n",
"promising results in enhancing model reasoning abilities through RL and distillation.\n"
"\n",
"The conclusions in this document are:\n",
"\n",
"- DeepSeek-R1-Zero, a pure reinforcement learning (RL) approach without cold-start data, achieves\n",
"strong performance across various tasks.\n",
"- DeepSeek-R1, which combines cold-start data with iterative RL fine-tuning, is more powerful and\n",
"achieves performance comparable to OpenAI-o1-1217 on a range of tasks.\n",
"- Distilling DeepSeek-R1’s reasoning capabilities into smaller dense models is promising; for\n",
"example, DeepSeek-R1-Distill-Qwen-1.5B outperforms GPT-4o and Claude-3.5-Sonnet on math benchmarks,\n",
"and other dense models also show significant improvements over similar instruction-tuned models.\n",
"\n",
"These results demonstrate the effectiveness of the RL-based approach and the potential for\n",
"distilling reasoning abilities into smaller models.\n"
]
}
],
@ -531,7 +543,7 @@
"Provide a clear, concise answer based only on the context provided.\n",
"\"\"\"\n",
"\n",
"print('Generated Answer:')\n",
"print('Generated Answer:\\n')\n",
"answer = await call_llm(answer_prompt)\n",
"show(answer)"
]