Skip to content

Commit 0914d81

Browse files
authored
Merge pull request #23 from cnblogs/upgrade-kernel-memory
feat: upgrade kernel memory
2 parents ae525cf + 05e9301 commit 0914d81

File tree

8 files changed

+44
-11
lines changed

8 files changed

+44
-11
lines changed

Diff for: src/KernelMemory.DashScope/DashScopeTextEmbeddingGenerator.cs

+6
Original file line numberDiff line numberDiff line change
@@ -24,6 +24,12 @@ public int CountTokens(string text)
2424
return tokenizer?.CountTokens(text) ?? text.Length;
2525
}
2626

27+
/// <inheritdoc />
28+
public IReadOnlyList<string> GetTokens(string text)
29+
{
30+
return tokenizer?.GetTokens(text) ?? [text];
31+
}
32+
2733
/// <inheritdoc />
2834
public async Task<Embedding> GenerateEmbeddingAsync(
2935
string text,

Diff for: src/KernelMemory.DashScope/DashScopeTextGenerator.cs

+6
Original file line numberDiff line numberDiff line change
@@ -30,6 +30,12 @@ public int CountTokens(string text)
3030
return tokenizer?.CountTokens(text) ?? QWenTokenizer.CountTokensStatic(text);
3131
}
3232

33+
/// <inheritdoc />
34+
public IReadOnlyList<string> GetTokens(string text)
35+
{
36+
return tokenizer?.GetTokens(text) ?? QWenTokenizer.GetTokensStatic(text);
37+
}
38+
3339
/// <inheritdoc />
3440
public async IAsyncEnumerable<string> GenerateTextAsync(
3541
string prompt,

Diff for: src/KernelMemory.DashScope/KernelMemory.DashScope.csproj

+2-2
Original file line numberDiff line numberDiff line change
@@ -19,8 +19,8 @@
1919

2020
<ItemGroup>
2121
<PackageReference Include="Microsoft.DeepDev.TokenizerLib" Version="1.3.3" />
22-
<PackageReference Include="Microsoft.KernelMemory.Abstractions" Version="0.65.240620.1" />
23-
<PackageReference Include="Cnblogs.DashScope.Core" Version="0.2.2" />
22+
<PackageReference Include="Microsoft.KernelMemory.Abstractions" Version="0.70.240803.1" />
23+
<PackageReference Include="Cnblogs.DashScope.Core" Version="0.3.0" />
2424
</ItemGroup>
2525

2626
<ItemGroup>

Diff for: src/KernelMemory.DashScope/LengthTokenizer.cs

+6
Original file line numberDiff line numberDiff line change
@@ -12,4 +12,10 @@ public int CountTokens(string text)
1212
{
1313
return text.Length;
1414
}
15+
16+
/// <inheritdoc />
17+
public IReadOnlyList<string> GetTokens(string text)
18+
{
19+
return text.Select(x => $"{x}").ToList();
20+
}
1521
}

Diff for: src/KernelMemory.DashScope/QWenTokenizer.cs

+16-1
Original file line numberDiff line numberDiff line change
@@ -13,7 +13,6 @@ public class QWenTokenizer : ITextTokenizer
1313
.Concat(Enumerable.Range(0, 205).Select(x => $"<|extra_{x}|>"))
1414
.Select((x, i) => new KeyValuePair<string, int>(x, 151643 + i))
1515
.ToDictionary();
16-
1716
private static readonly ITokenizer Tokenizer = TokenizerBuilder.CreateTokenizer(
1817
DashScopeEmbeddedResource.ReadBpeFile(),
1918
SpecialTokens,
@@ -45,6 +44,12 @@ public int CountTokens(string text)
4544
return Tokenizer.Encode(text).Count;
4645
}
4746

47+
/// <inheritdoc />
48+
public IReadOnlyList<string> GetTokens(string text)
49+
{
50+
return Tokenizer.Encode(text).Select(x => Tokenizer.Decode([x])).ToList();
51+
}
52+
4853
/// <summary>
4954
/// Count tokens.
5055
/// </summary>
@@ -54,4 +59,14 @@ public static int CountTokensStatic(string text)
5459
{
5560
return Tokenizer.Encode(text).Count;
5661
}
62+
63+
/// <summary>
64+
/// Get tokens
65+
/// </summary>
66+
/// <param name="text">The text to tokenizers.</param>
67+
/// <returns></returns>
68+
public static IReadOnlyList<string> GetTokensStatic(string text)
69+
{
70+
return Tokenizer.Encode(text).Select(x => Tokenizer.Decode([x])).ToList();
71+
}
5772
}

Diff for: src/SemanticKernel.DashScope/SemanticKernel.DashScope.csproj

+3-3
Original file line numberDiff line numberDiff line change
@@ -19,9 +19,9 @@
1919

2020
<ItemGroup>
2121
<FrameworkReference Include="Microsoft.AspNetCore.App"/>
22-
<PackageReference Include="Cnblogs.DashScope.Core" Version="0.2.2" />
23-
<PackageReference Include="JsonSchema.Net.Generation" Version="4.3.0.2" />
24-
<PackageReference Include="Microsoft.SemanticKernel.Core" Version="1.15.1" />
22+
<PackageReference Include="Cnblogs.DashScope.Core" Version="0.3.0" />
23+
<PackageReference Include="JsonSchema.Net.Generation" Version="4.5.0" />
24+
<PackageReference Include="Microsoft.SemanticKernel.Core" Version="1.16.2" />
2525
</ItemGroup>
2626

2727
<ItemGroup>

Diff for: test/KernelMemory.DashScope.UnitTests/KernelMemory.DashScope.UnitTests.csproj

+3-3
Original file line numberDiff line numberDiff line change
@@ -14,10 +14,10 @@
1414
<PrivateAssets>all</PrivateAssets>
1515
<IncludeAssets>runtime; build; native; contentfiles; analyzers; buildtransitive</IncludeAssets>
1616
</PackageReference>
17-
<PackageReference Include="Microsoft.KernelMemory.Core" Version="0.65.240620.1" />
17+
<PackageReference Include="Microsoft.KernelMemory.Core" Version="0.70.240803.1" />
1818
<PackageReference Include="Microsoft.NET.Test.Sdk" Version="17.10.0" />
19-
<PackageReference Include="xunit" Version="2.8.1" />
20-
<PackageReference Include="xunit.runner.visualstudio" Version="2.8.1">
19+
<PackageReference Include="xunit" Version="2.9.0" />
20+
<PackageReference Include="xunit.runner.visualstudio" Version="2.8.2">
2121
<PrivateAssets>all</PrivateAssets>
2222
<IncludeAssets>runtime; build; native; contentfiles; analyzers; buildtransitive</IncludeAssets>
2323
</PackageReference>

Diff for: test/SemanticKernel.DashScope.UnitTest/SemanticKernel.DashScope.UnitTest.csproj

+2-2
Original file line numberDiff line numberDiff line change
@@ -5,8 +5,8 @@
55
<IncludeAssets>runtime; build; native; contentfiles; analyzers; buildtransitive</IncludeAssets>
66
</PackageReference>
77
<PackageReference Include="Microsoft.NET.Test.Sdk" Version="17.10.0" />
8-
<PackageReference Include="xunit" Version="2.8.1" />
9-
<PackageReference Include="xunit.runner.visualstudio" Version="2.8.1">
8+
<PackageReference Include="xunit" Version="2.9.0" />
9+
<PackageReference Include="xunit.runner.visualstudio" Version="2.8.2">
1010
<PrivateAssets>all</PrivateAssets>
1111
<IncludeAssets>runtime; build; native; contentfiles; analyzers; buildtransitive</IncludeAssets>
1212
</PackageReference>

0 commit comments

Comments
 (0)