Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

feat:Updated Deep Infra models #165

Merged
merged 1 commit into from
Feb 17, 2025
Merged

feat:Updated Deep Infra models #165

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
30 changes: 15 additions & 15 deletions src/DeepInfra/src/DeepInfraModelIds.cs
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -9,7 +9,7 @@ public enum DeepInfraModelIds
/// <summary>
/// Name: DeepSeek-R1 <br/>
/// Organization: deepseek-ai <br/>
/// Context Length: 32768 <br/>
/// Context Length: 65536 <br/>
/// Prompt Cost: $0.75/MTok <br/>
/// Completion Cost: $0.75/MTok <br/>
/// Description: We introduce DeepSeek-R1, which incorporates cold-start data before RL. DeepSeek-R1 achieves performance comparable to OpenAI-o1 across math, code, and reasoning tasks. <br/>
Expand All @@ -31,7 +31,7 @@ public enum DeepInfraModelIds
/// <summary>
/// Name: DeepSeek-V3 <br/>
/// Organization: deepseek-ai <br/>
/// Context Length: 32768 <br/>
/// Context Length: 65536 <br/>
/// Prompt Cost: $0.49/MTok <br/>
/// Completion Cost: $0.49/MTok <br/>
/// Description: DeepSeek-V3, a strong Mixture-of-Experts (MoE) language model with 671B total parameters with 37B activated for each token. To achieve efficient inference and cost-effective training, DeepSeek-V3 adopts Multi-head Latent Attention (MLA) and DeepSeekMoE architectures, which were thoroughly validated in DeepSeek-V2. <br/>
Expand Down Expand Up @@ -127,17 +127,6 @@ public enum DeepInfraModelIds
/// </summary>
MetaLlama31405BInstruct,

/// <summary>
/// Name: QwQ-32B-Preview <br/>
/// Organization: Qwen <br/>
/// Context Length: 32768 <br/>
/// Prompt Cost: $0.12/MTok <br/>
/// Completion Cost: $0.12/MTok <br/>
/// Description: QwQ is an experimental research model developed by the Qwen Team, designed to advance AI reasoning capabilities. This model embodies the spirit of philosophical inquiry, approaching problems with genuine wonder and doubt. QwQ demonstrates impressive analytical abilities, achieving scores of 65.2% on GPQA, 50.0% on AIME, 90.6% on MATH-500, and 50.0% on LiveCodeBench. With its contemplative approach and exceptional performance on complex problems. <br/>
/// HuggingFace Url: <a href="https://huggingface.co/Qwen/QwQ-32B-Preview">https://huggingface.co/Qwen/QwQ-32B-Preview</a>
/// </summary>
Qwq32BPreview,

/// <summary>
/// Name: Meta-Llama-3.1-8B-Instruct-Turbo <br/>
/// Organization: meta-llama <br/>
Expand Down Expand Up @@ -186,8 +175,8 @@ public enum DeepInfraModelIds
/// Name: Qwen2.5-72B-Instruct <br/>
/// Organization: Qwen <br/>
/// Context Length: 32768 <br/>
/// Prompt Cost: $0.23/MTok <br/>
/// Completion Cost: $0.23/MTok <br/>
/// Prompt Cost: $0.13/MTok <br/>
/// Completion Cost: $0.13/MTok <br/>
/// Description: Qwen2.5 is a model pretrained on a large-scale dataset of up to 18 trillion tokens, offering significant improvements in knowledge, coding, mathematics, and instruction following compared to its predecessor Qwen2. The model also features enhanced capabilities in generating long texts, understanding structured data, and generating structured outputs, while supporting multilingual capabilities for over 29 languages. <br/>
/// HuggingFace Url: <a href="https://huggingface.co/Qwen/Qwen2.5-72B-Instruct">https://huggingface.co/Qwen/Qwen2.5-72B-Instruct</a>
/// </summary>
Expand Down Expand Up @@ -336,6 +325,17 @@ public enum DeepInfraModelIds
/// </summary>
Qvq72BPreview,

/// <summary>
/// Name: QwQ-32B-Preview <br/>
/// Organization: Qwen <br/>
/// Context Length: 32768 <br/>
/// Prompt Cost: $0.12/MTok <br/>
/// Completion Cost: $0.12/MTok <br/>
/// Description: QwQ is an experimental research model developed by the Qwen Team, designed to advance AI reasoning capabilities. This model embodies the spirit of philosophical inquiry, approaching problems with genuine wonder and doubt. QwQ demonstrates impressive analytical abilities, achieving scores of 65.2% on GPQA, 50.0% on AIME, 90.6% on MATH-500, and 50.0% on LiveCodeBench. With its contemplative approach and exceptional performance on complex problems. <br/>
/// HuggingFace Url: <a href="https://huggingface.co/Qwen/QwQ-32B-Preview">https://huggingface.co/Qwen/QwQ-32B-Preview</a>
/// </summary>
Qwq32BPreview,

/// <summary>
/// Name: Qwen2-72B-Instruct <br/>
/// Organization: Qwen <br/>
Expand Down
8 changes: 4 additions & 4 deletions src/DeepInfra/src/DeepInfraModelProvider.cs
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -9,9 +9,9 @@ public static class DeepInfraModelProvider
{
private static Dictionary<DeepInfraModelIds, ChatModelMetadata> Models { get; set; } = new()
{
{ DeepInfraModelIds.DeepseekR1, ToMetadata("deepseek-ai/DeepSeek-R1",32768,7.5E-07,2.4E-06)},
{ DeepInfraModelIds.DeepseekR1, ToMetadata("deepseek-ai/DeepSeek-R1",65536,7.5E-07,2.4E-06)},
{ DeepInfraModelIds.DeepseekR1DistillLlama70B, ToMetadata("deepseek-ai/DeepSeek-R1-Distill-Llama-70B",131072,2.3000000000000002E-07,6.9E-07)},
{ DeepInfraModelIds.DeepseekV3, ToMetadata("deepseek-ai/DeepSeek-V3",32768,4.9E-07,8.900000000000001E-07)},
{ DeepInfraModelIds.DeepseekV3, ToMetadata("deepseek-ai/DeepSeek-V3",65536,4.9E-07,8.900000000000001E-07)},
{ DeepInfraModelIds.Llama3370BInstructTurbo, ToMetadata("meta-llama/Llama-3.3-70B-Instruct-Turbo",131072,1.2E-07,3E-07)},
{ DeepInfraModelIds.Llama3370BInstruct, ToMetadata("meta-llama/Llama-3.3-70B-Instruct",131072,2.3000000000000002E-07,4.0000000000000003E-07)},
{ DeepInfraModelIds.MistralSmall24BInstruct2501, ToMetadata("mistralai/Mistral-Small-24B-Instruct-2501",32768,7E-08,1.4E-07)},
Expand All @@ -20,12 +20,11 @@ public static class DeepInfraModelProvider
{ DeepInfraModelIds.MetaLlama3170BInstruct, ToMetadata("meta-llama/Meta-Llama-3.1-70B-Instruct",131072,2.3000000000000002E-07,4.0000000000000003E-07)},
{ DeepInfraModelIds.MetaLlama318BInstruct, ToMetadata("meta-llama/Meta-Llama-3.1-8B-Instruct",131072,3E-08,5.0000000000000004E-08)},
{ DeepInfraModelIds.MetaLlama31405BInstruct, ToMetadata("meta-llama/Meta-Llama-3.1-405B-Instruct",32768,8.000000000000001E-07,8.000000000000001E-07)},
{ DeepInfraModelIds.Qwq32BPreview, ToMetadata("Qwen/QwQ-32B-Preview",32768,1.2E-07,1.8E-07)},
{ DeepInfraModelIds.MetaLlama318BInstructTurbo, ToMetadata("meta-llama/Meta-Llama-3.1-8B-Instruct-Turbo",131072,2E-08,5.0000000000000004E-08)},
{ DeepInfraModelIds.MetaLlama3170BInstructTurbo, ToMetadata("meta-llama/Meta-Llama-3.1-70B-Instruct-Turbo",131072,1.2E-07,3E-07)},
{ DeepInfraModelIds.Qwen25Coder32BInstruct, ToMetadata("Qwen/Qwen2.5-Coder-32B-Instruct",32768,7E-08,1.6E-07)},
{ DeepInfraModelIds.Llama31Nemotron70BInstruct, ToMetadata("nvidia/Llama-3.1-Nemotron-70B-Instruct",131072,1.2E-07,3E-07)},
{ DeepInfraModelIds.Qwen2572BInstruct, ToMetadata("Qwen/Qwen2.5-72B-Instruct",32768,2.3000000000000002E-07,4.0000000000000003E-07)},
{ DeepInfraModelIds.Qwen2572BInstruct, ToMetadata("Qwen/Qwen2.5-72B-Instruct",32768,1.3E-07,4.0000000000000003E-07)},
{ DeepInfraModelIds.Llama3290BVisionInstruct, ToMetadata("meta-llama/Llama-3.2-90B-Vision-Instruct",32768,3.5E-07,4.0000000000000003E-07)},
{ DeepInfraModelIds.Llama3211BVisionInstruct, ToMetadata("meta-llama/Llama-3.2-11B-Vision-Instruct",131072,6E-08,6E-08)},
{ DeepInfraModelIds.Wizardlm28X22b, ToMetadata("microsoft/WizardLM-2-8x22B",65536,5E-07,5E-07)},
Expand All @@ -39,6 +38,7 @@ public static class DeepInfraModelProvider
{ DeepInfraModelIds.SkyT132BPreview, ToMetadata("NovaSky-AI/Sky-T1-32B-Preview",32768,1.2E-07,1.8E-07)},
{ DeepInfraModelIds.PhindCodellama34BV2, ToMetadata("Phind/Phind-CodeLlama-34B-v2",4096,6E-07,6E-07)},
{ DeepInfraModelIds.Qvq72BPreview, ToMetadata("Qwen/QVQ-72B-Preview",32000,2.5E-07,5E-07)},
{ DeepInfraModelIds.Qwq32BPreview, ToMetadata("Qwen/QwQ-32B-Preview",32768,1.2E-07,1.8E-07)},
{ DeepInfraModelIds.Qwen272BInstruct, ToMetadata("Qwen/Qwen2-72B-Instruct",32768,3.5E-07,4.0000000000000003E-07)},
{ DeepInfraModelIds.Qwen27BInstruct, ToMetadata("Qwen/Qwen2-7B-Instruct",32768,6E-08,6E-08)},
{ DeepInfraModelIds.Qwen257BInstruct, ToMetadata("Qwen/Qwen2.5-7B-Instruct",32768,2E-08,5.0000000000000004E-08)},
Expand Down
10 changes: 5 additions & 5 deletions src/DeepInfra/src/Predefined/AllModels.cs
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -55,11 +55,6 @@ public class MetaLlama318BInstructModel(DeepInfraProvider provider) : DeepInfraM
public class MetaLlama31405BInstructModel(DeepInfraProvider provider) : DeepInfraModel(provider, DeepInfraModelIds.MetaLlama31405BInstruct);


/// <inheritdoc cref="DeepInfraModelIds.Qwq32BPreview"/>
/// <param name="provider">Deep Infra Provider Instance</param>
public class Qwq32BPreviewModel(DeepInfraProvider provider) : DeepInfraModel(provider, DeepInfraModelIds.Qwq32BPreview);


/// <inheritdoc cref="DeepInfraModelIds.MetaLlama318BInstructTurbo"/>
/// <param name="provider">Deep Infra Provider Instance</param>
public class MetaLlama318BInstructTurboModel(DeepInfraProvider provider) : DeepInfraModel(provider, DeepInfraModelIds.MetaLlama318BInstructTurbo);
Expand Down Expand Up @@ -150,6 +145,11 @@ public class PhindCodellama34BV2Model(DeepInfraProvider provider) : DeepInfraMod
public class Qvq72BPreviewModel(DeepInfraProvider provider) : DeepInfraModel(provider, DeepInfraModelIds.Qvq72BPreview);


/// <inheritdoc cref="DeepInfraModelIds.Qwq32BPreview"/>
/// <param name="provider">Deep Infra Provider Instance</param>
public class Qwq32BPreviewModel(DeepInfraProvider provider) : DeepInfraModel(provider, DeepInfraModelIds.Qwq32BPreview);


/// <inheritdoc cref="DeepInfraModelIds.Qwen272BInstruct"/>
/// <param name="provider">Deep Infra Provider Instance</param>
public class Qwen272BInstructModel(DeepInfraProvider provider) : DeepInfraModel(provider, DeepInfraModelIds.Qwen272BInstruct);
Expand Down
Loading