我的目标是使用google模型适应来提高语音到文本的准确性，但是这些api在任何地方都没有很好的记录。
https://cloud.google.com/speech-to-text/docs/reference/rest/v1p1beta1/projects.locations.customClasses
我尝试创建一个具有200000个值的自定义类。在这个计数之上，它给出了有效载荷大小的错误，而不是条目计数限制。我在哪里可以找到API及其限制的适当信息/详细信息。
我正在使用Ruby库创建自定义类。
创建自定义类的代码。
cname = "TestClass"
items = 3_00_000.times.map{|e| Google::Cloud::Speech::V1p1beta1::CustomClass::ClassItem.new(value: Faker::Name.name) };
_class = Google::Cloud::Speech::V1p1beta1::CustomClass.new(name: cname, items: items);
request = Google::Cloud::Speech::V1p1beta1::CreateCustomClassRequest.new({custom_class: _class, parent: "projects/<projectID>/locations/global", custom_class_id: cname})
_klass = client.create_custom_class request
得到下面的错误看起来像是用10_000_000的值创建/更新。
Google::Cloud::InvalidArgumentError: 3:Request payload size exceeds the limit: 10485760 bytes.. debug_error_string:{"created":"@1628230030.306827000","description":"Error received from peer ipv4:142.251.42.10:443","file":"src/core/lib/surface/call.cc","file_line":1067,"grpc_message":"Request payload size exceeds the limit: 10485760 bytes.","grpc_status":3}
这是所有关于API的公开文档。
https://cloud.google.com/speech/docs/
https://cloud.google.com/speech-to-text/docs/release-notes
https://cloud.google.com/speech-to-text/pricing
https://cloud.google.com/speech-to-text/quotas
https://cloud.google.com/speech-to-text/sla
https://cloud.google.com/speech-to-text/docs/support#troubleshooting
https://cloud.google.com/speech-to-text/docs/best-practices
https://cloud.google.com/speech-to-text/docs/encoding
https://cloud.google.com/speech-to-text/docs/languages
https://cloud.google.com/speech-to-text/docs/apis
https://cloud.google.com/speech-to-text/docs/concepts
https://cloud.google.com/speech-to-text/docs/how-to
https://cloud.google.com/speech/docs/tutorials