AIGC ..... 4 1.2 AGI ..... 5 1.3 大模型 ..... 5 1.4 基础概念 ..... 6 1.4.1 上下文窗口 ..... 6 1.4.2 单位 B 和 T ..... 6 2 AI 工具梳理 ..... 6 2.1 问答 ..... 6 2.1.1 ChatGPT ..... 6 2.1.2 Claude ..... 7 2 BERT。这些模型因其高效的学习能力和强大的通用性而受到关注。 开源大模型以 Meta 的 Llama 系列，2024 年 4 月，Llama3 发布，包括 8B 和 70B 模型。 图 2，时间线主要根据技术论文的发布日期（例如提交至 arXiv 的日期）来确定大型语言模型（大小超过 10B）的发展历程。如果没有相应的论文，我们将模型的日期设定为其公开发布或宣布的最早时间。我们用黄色标记那些公开可用的模型检查点。由于空间 #### 1.4.2 单位 B 和 T 在 AI 大模型中，常用的两个单位是 B 和 T。 B（十亿，Billion）：在英文里是 Billion 的缩写，表示十亿。对于 AI 大模型来说，B 一般用于描述模型的参数数量。例如，具有 50B 参数的模型代表这个模型有 50 亿个参数。Ollama3 有尺寸 8B 和 70B，Phi-3-mini 有 3.8B 参数等。 T（万亿，Trillion）：在英文里是

# Falcon Documentation Release 3.0.0b2 Kurt Griffiths et al. Mar 03, 2021 Powered by TCPDF (www.tcpdf.org) ## CONTENTS 1 What People are Saying 3 2 Quick Links 5 3 Features 7 4 Who's Using 5.5.10 Changelog for Falcon 0.2.0 324 Python Module Index 327 Index 329 ### Release v3.0b2 (Installation) Falcon is a minimalist WSGI library for building speedy web APIs and app backends. '123e4567-e89b-12d3-a456-426655440000' monkeypatch.setattr('uuid.uuid4', lambda: fake_uuid) # When the service receives an image through POST... fake_image_bytes = b'fake-image-bytes'

# Falcon Documentation Release 3.1.1b3 Kurt Griffiths et al. Nov 11, 2022 Powered by TCPDF (www.tcpdf.org) ## CONTENTS 1 Quick Links 3 2 What People are Saying 5 3 Features 7 4 Who's Using 5.5.13 Changelog for Falcon 0.2.0 361 Python Module Index 365 Index 367 ### Release v3.1b3 (Installation) Falcon is a minimalist ASGI/WSGI framework for building mission-critical REST APIs mock_open() monkeypatch.setattr('io.open', mock_file_open) fake_uuid = '123e4567-e89b-12d3-a456-426655440000' monkeypatch.setattr('uuid.uuid4', lambda: fake_uuid) (continues

# The Falcon Web Framework Release v3.1b3 (Installation) Falcon [https://falconframework.org] is a minimalist ASGI/WSGI framework for building mission-critical REST APIs and microservices, with a focus '123e4567-e89b-12d3-a456-426655440000' monkeypatch.setattr('uuid.uuid4', lambda: fake_uuid) # When the service receives an image through POST... fake_image_bytes = b'fake-image-bytes' image_content_type = 'image/xyz' response = client.simulate_post( '/images', body=b'some-fake-bytes', headers={'content-type': image_content_type} ) assert response

# Falcon Documentation Release 3.1.1b1 Kurt Griffiths et al. Nov 09, 2022 Powered by TCPDF (www.tcpdf.org) ## CONTENTS 1 Quick Links 3 2 What People are Saying 5 3 Features 7 4 Who's Using 5.5.13 Changelog for Falcon 0.2.0 361 Python Module Index 365 Index 367 ### Release v3.1b1 (Installation) Falcon is a minimalist ASGI/WSGI framework for building mission-critical REST APIs mock_open() monkeypatch.setattr('io.open', mock_file_open) fake_uuid = '123e4567-e89b-12d3-a456-426655440000' monkeypatch.setattr('uuid.uuid4', lambda: fake_uuid) (continues

class QuoteResource: # The Falcon Web Framework Release v3.0b1 (Installation) Falcon is a minimalist WSGI library for building speedy web APIs and app backends. We like to think of Falcon as the Dieter '123e4567-e89b-12d3-a456-426655440000' monkeypatch.setattr('uuid.uuid4', lambda: fake_uuid) # When the service receives an image through POST... fake_image_bytes = b'fake-image-bytes' image_content_type = 'image/xyz' response = client.simulate_post( '/images', body=b'some-fake-bytes', headers={'content-type': image_content_type} ) assert response

# Falcon Documentation Release 3.0.0b1 Kurt Griffiths et al. Dec 23, 2020 Powered by TCPDF (www.tcpdf.org) ## CONTENTS 1 What People are Saying 3 2 Quick Links 5 3 Features 7 4 Who's Using 5.5.10 Changelog for Falcon 0.2.0 322 Python Module Index 325 Index 327 ### Release v3.0b1 (Installation) Falcon is a minimalist WSGI library for building speedy web APIs and app backends. '123e4567-e89b-12d3-a456-426655440000' monkeypatch.setattr('uuid.uuid4', lambda: fake_uuid) # When the service receives an image through POST... fake_image_bytes = b'fake-image-bytes'

class QuoteResource: # The Falcon Web Framework Release v3.0b2 (Installation) Falcon is a minimalist WSGI library for building speedy web APIs and app backends. We like to think of Falcon as the Dieter '123e4567-e89b-12d3-a456-426655440000' monkeypatch.setattr('uuid.uuid4', lambda: fake_uuid) # When the service receives an image through POST... fake_image_bytes = b'fake-image-bytes' image_content_type = 'image/xyz' response = client.simulate_post( '/images', body=b'some-fake-bytes', headers={'content-type': image_content_type} ) assert response

# The Falcon Web Framework Release v3.1b1 (Installation) Falcon [https://falconframework.org] is a minimalist ASGI/WSGI framework for building mission-critical REST APIs and microservices, with a focus '123e4567-e89b-12d3-a456-426655440000' monkeypatch.setattr('uuid.uuid4', lambda: fake_uuid) # When the service receives an image through POST... fake_image_bytes = b'fake-image-bytes' image_content_type = 'image/xyz' response = client.simulate_post( '/images', body=b'some-fake-bytes', headers={'content-type': image_content_type} ) assert response

靳宇栋（Krahets）  Release 1.0.0b1 2023-03-01 ## 序 两年前，我在力扣上分享了《剑指 Offer》系列题解，受到了很多小伙伴的喜爱与支持。在此期间，我也回复了许多读者的评论问题，遇到最多的问题是“如何入门学习算法”。我渐渐也对这个问题好奇了起来。 实际测试 假设我们现在有算法 A 和算法 B，都能够解决同一问题，现在需要对比两个算法之间的效率。我们能够想到的最直接的方式，就是找一台计算机，把两个算法都完整跑一遍，并监控记录运行时间和内存占用情况。这种评估方式能够反映真实情况，但是也存在很大的硬伤。 难以排除测试环境的干扰因素。硬件配置会影响到算法的性能表现。例如，在某台计算机中，算法 A 比算法 B 运行时间更短；但换到另一台配置不同的计算 。这意味着我们需要在各种机器上展开测试，而这是不现实的。 展开完整测试非常耗费资源。随着输入数据量的大小变化，算法会呈现出不同的效率表现。比如，有可能输入数据量较小时，算法 A 运行时间短于算法 B，而在输入数据量较大时，测试结果截然相反。因此，若想要达到具有说服力的对比结果，那么需要输入各种体量数据，这样的测试需要占用大量计算资源。 ## 理论估算 既然实际测试具有很大的局限性，那么我们是