[go: nahoru, domu]
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

Jump to bottom

Select the best device in examples #507

Merged
merged 1 commit into from
Jul 1, 2024
Merged

Select the best device in examples #507

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
Select the best device in examples
  • Loading branch information
@EricLBuehler
EricLBuehler committed Jul 1, 2024
commit 8666d06fa217980cc3e51300dc5beb54c6c6d57b
16 changes: 14 additions & 2 deletions mistralrs/examples/gemma2/main.rs
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -4,9 +4,21 @@ use tokio::sync::mpsc::channel;
use mistralrs::{
Constraint, Device, DeviceMapMetadata, MistralRs, MistralRsBuilder, ModelDType,
NormalLoaderBuilder, NormalLoaderType, NormalRequest, NormalSpecificConfig, Request,
RequestMessage, Response, SamplingParams, SchedulerMethod, TokenSource,
RequestMessage, Response, Result, SamplingParams, SchedulerMethod, TokenSource,
};

/// Gets the best device, cpu, cuda if compiled with CUDA
pub(crate) fn best_device() -> Result<Device> {
#[cfg(not(feature = "metal"))]
{
Device::cuda_if_available(0)
}
#[cfg(feature = "metal")]
{
Device::new_metal(0)
}
}

fn setup() -> anyhow::Result<Arc<MistralRs>> {
// Select a Mistral model
let loader = NormalLoaderBuilder::new(
Expand All @@ -24,7 +36,7 @@ fn setup() -> anyhow::Result<Arc<MistralRs>> {
None,
TokenSource::CacheToken,
&ModelDType::Auto,
&Device::cuda_if_available(0)?,
&best_device()?,
false,
DeviceMapMetadata::dummy(),
None,
Expand Down
18 changes: 15 additions & 3 deletions mistralrs/examples/gguf_locally/main.rs
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -3,10 +3,22 @@ use tokio::sync::mpsc::channel;

use mistralrs::{
Constraint, Device, DeviceMapMetadata, GGUFLoaderBuilder, GGUFSpecificConfig, MistralRs,
MistralRsBuilder, ModelDType, NormalRequest, Request, RequestMessage, Response, SamplingParams,
SchedulerMethod, TokenSource,
MistralRsBuilder, ModelDType, NormalRequest, Request, RequestMessage, Response, Result,
SamplingParams, SchedulerMethod, TokenSource,
};

/// Gets the best device, cpu, cuda if compiled with CUDA
pub(crate) fn best_device() -> Result<Device> {
#[cfg(not(feature = "metal"))]
{
Device::cuda_if_available(0)
}
#[cfg(feature = "metal")]
{
Device::new_metal(0)
}
}

fn setup() -> anyhow::Result<Arc<MistralRs>> {
// Select a Mistral model
// We do not use any files from HF servers here, and instead load the
Expand All @@ -25,7 +37,7 @@ fn setup() -> anyhow::Result<Arc<MistralRs>> {
None,
TokenSource::CacheToken,
&ModelDType::Auto,
&Device::cuda_if_available(0)?,
&best_device()?,
false,
DeviceMapMetadata::dummy(),
None,
Expand Down
16 changes: 14 additions & 2 deletions mistralrs/examples/grammar/main.rs
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -4,9 +4,21 @@ use tokio::sync::mpsc::channel;
use mistralrs::{
Constraint, Device, DeviceMapMetadata, MistralRs, MistralRsBuilder, ModelDType,
NormalLoaderBuilder, NormalLoaderType, NormalRequest, NormalSpecificConfig, Request,
RequestMessage, Response, SamplingParams, SchedulerMethod, TokenSource,
RequestMessage, Response, Result, SamplingParams, SchedulerMethod, TokenSource,
};

/// Gets the best device, cpu, cuda if compiled with CUDA
pub(crate) fn best_device() -> Result<Device> {
#[cfg(not(feature = "metal"))]
{
Device::cuda_if_available(0)
}
#[cfg(feature = "metal")]
{
Device::new_metal(0)
}
}

fn setup() -> anyhow::Result<Arc<MistralRs>> {
// Select a Mistral model
let loader = NormalLoaderBuilder::new(
Expand All @@ -24,7 +36,7 @@ fn setup() -> anyhow::Result<Arc<MistralRs>> {
None,
TokenSource::CacheToken,
&ModelDType::Auto,
&Device::cuda_if_available(0)?,
&best_device()?,
false,
DeviceMapMetadata::dummy(),
None,
Expand Down
16 changes: 14 additions & 2 deletions mistralrs/examples/idefics2/main.rs
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -6,10 +6,22 @@ use tokio::sync::mpsc::channel;

use mistralrs::{
Constraint, Device, DeviceMapMetadata, MistralRs, MistralRsBuilder, ModelDType, NormalRequest,
Request, RequestMessage, Response, SamplingParams, SchedulerMethod, TokenSource,
Request, RequestMessage, Response, Result, SamplingParams, SchedulerMethod, TokenSource,
VisionLoaderBuilder, VisionLoaderType, VisionSpecificConfig,
};

/// Gets the best device, cpu, cuda if compiled with CUDA
pub(crate) fn best_device() -> Result<Device> {
#[cfg(not(feature = "metal"))]
{
Device::cuda_if_available(0)
}
#[cfg(feature = "metal")]
{
Device::new_metal(0)
}
}

fn setup() -> anyhow::Result<Arc<MistralRs>> {
// Select a Mistral model
let loader = VisionLoaderBuilder::new(
Expand All @@ -27,7 +39,7 @@ fn setup() -> anyhow::Result<Arc<MistralRs>> {
None,
TokenSource::CacheToken,
&ModelDType::Auto,
&Device::cuda_if_available(0)?,
&best_device()?,
false,
DeviceMapMetadata::dummy(),
None,
Expand Down
16 changes: 14 additions & 2 deletions mistralrs/examples/isq/main.rs
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -4,9 +4,21 @@ use tokio::sync::mpsc::channel;
use mistralrs::{
Constraint, Device, DeviceMapMetadata, GgmlDType, MistralRs, MistralRsBuilder, ModelDType,
NormalLoaderBuilder, NormalLoaderType, NormalRequest, NormalSpecificConfig, Request,
RequestMessage, Response, SamplingParams, SchedulerMethod, TokenSource,
RequestMessage, Response, Result, SamplingParams, SchedulerMethod, TokenSource,
};

/// Gets the best device, cpu, cuda if compiled with CUDA
pub(crate) fn best_device() -> Result<Device> {
#[cfg(not(feature = "metal"))]
{
Device::cuda_if_available(0)
}
#[cfg(feature = "metal")]
{
Device::new_metal(0)
}
}

fn setup() -> anyhow::Result<Arc<MistralRs>> {
// Select a Mistral model
let loader = NormalLoaderBuilder::new(
Expand All @@ -24,7 +36,7 @@ fn setup() -> anyhow::Result<Arc<MistralRs>> {
None,
TokenSource::CacheToken,
&ModelDType::Auto,
&Device::cuda_if_available(0)?,
&best_device()?,
false,
DeviceMapMetadata::dummy(),
Some(GgmlDType::Q4K), // In-situ quantize the model into q4k
Expand Down
16 changes: 14 additions & 2 deletions mistralrs/examples/lora/main.rs
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -4,9 +4,21 @@ use tokio::sync::mpsc::channel;
use mistralrs::{
Constraint, Device, DeviceMapMetadata, MistralRs, MistralRsBuilder, ModelDType,
NormalLoaderBuilder, NormalLoaderType, NormalRequest, NormalSpecificConfig, Request,
RequestMessage, Response, SamplingParams, SchedulerMethod, TokenSource,
RequestMessage, Response, Result, SamplingParams, SchedulerMethod, TokenSource,
};

/// Gets the best device, cpu, cuda if compiled with CUDA
pub(crate) fn best_device() -> Result<Device> {
#[cfg(not(feature = "metal"))]
{
Device::cuda_if_available(0)
}
#[cfg(feature = "metal")]
{
Device::new_metal(0)
}
}

fn setup() -> anyhow::Result<Arc<MistralRs>> {
// Select a Mistral model
let loader =
Expand All @@ -31,7 +43,7 @@ fn setup() -> anyhow::Result<Arc<MistralRs>> {
None,
TokenSource::CacheToken,
&ModelDType::Auto,
&Device::cuda_if_available(0)?,
&best_device()?,
false,
DeviceMapMetadata::dummy(),
None,
Expand Down
16 changes: 14 additions & 2 deletions mistralrs/examples/lora_activation/main.rs
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -4,9 +4,21 @@ use tokio::sync::mpsc::channel;
use mistralrs::{
Constraint, Device, DeviceMapMetadata, MistralRs, MistralRsBuilder, ModelDType,
NormalLoaderBuilder, NormalLoaderType, NormalRequest, NormalSpecificConfig, Request,
RequestMessage, Response, SamplingParams, SchedulerMethod, TokenSource,
RequestMessage, Response, Result, SamplingParams, SchedulerMethod, TokenSource,
};

/// Gets the best device, cpu, cuda if compiled with CUDA
pub(crate) fn best_device() -> Result<Device> {
#[cfg(not(feature = "metal"))]
{
Device::cuda_if_available(0)
}
#[cfg(feature = "metal")]
{
Device::new_metal(0)
}
}

fn setup() -> anyhow::Result<Arc<MistralRs>> {
// Select a Mistral model
let loader =
Expand All @@ -31,7 +43,7 @@ fn setup() -> anyhow::Result<Arc<MistralRs>> {
None,
TokenSource::CacheToken,
&ModelDType::Auto,
&Device::cuda_if_available(0)?,
&best_device()?,
false,
DeviceMapMetadata::dummy(),
None,
Expand Down
16 changes: 14 additions & 2 deletions mistralrs/examples/phi3v/main.rs
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -6,10 +6,22 @@ use tokio::sync::mpsc::channel;

use mistralrs::{
Constraint, Device, DeviceMapMetadata, MistralRs, MistralRsBuilder, ModelDType, NormalRequest,
Request, RequestMessage, Response, SamplingParams, SchedulerMethod, TokenSource,
Request, RequestMessage, Response, Result, SamplingParams, SchedulerMethod, TokenSource,
VisionLoaderBuilder, VisionLoaderType, VisionSpecificConfig,
};

/// Gets the best device, cpu, cuda if compiled with CUDA
pub(crate) fn best_device() -> Result<Device> {
#[cfg(not(feature = "metal"))]
{
Device::cuda_if_available(0)
}
#[cfg(feature = "metal")]
{
Device::new_metal(0)
}
}

fn setup() -> anyhow::Result<Arc<MistralRs>> {
// Select a Mistral model
let loader = VisionLoaderBuilder::new(
Expand All @@ -27,7 +39,7 @@ fn setup() -> anyhow::Result<Arc<MistralRs>> {
None,
TokenSource::CacheToken,
&ModelDType::Auto,
&Device::cuda_if_available(0)?,
&best_device()?,
false,
DeviceMapMetadata::dummy(),
None,
Expand Down
18 changes: 15 additions & 3 deletions mistralrs/examples/quantized/main.rs
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -3,10 +3,22 @@ use tokio::sync::mpsc::channel;

use mistralrs::{
Constraint, Device, DeviceMapMetadata, GGUFLoaderBuilder, GGUFSpecificConfig, MistralRs,
MistralRsBuilder, ModelDType, NormalRequest, Request, RequestMessage, Response, SamplingParams,
SchedulerMethod, TokenSource,
MistralRsBuilder, ModelDType, NormalRequest, Request, RequestMessage, Response, Result,
SamplingParams, SchedulerMethod, TokenSource,
};

/// Gets the best device, cpu, cuda if compiled with CUDA
pub(crate) fn best_device() -> Result<Device> {
#[cfg(not(feature = "metal"))]
{
Device::cuda_if_available(0)
}
#[cfg(feature = "metal")]
{
Device::new_metal(0)
}
}

fn setup() -> anyhow::Result<Arc<MistralRs>> {
// Select a Mistral model
// This uses a model, tokenizer, and chat template, from HF hub.
Expand All @@ -23,7 +35,7 @@ fn setup() -> anyhow::Result<Arc<MistralRs>> {
None,
TokenSource::CacheToken,
&ModelDType::default(),
&Device::cuda_if_available(0)?,
&best_device()?,
false,
DeviceMapMetadata::dummy(),
None,
Expand Down
16 changes: 14 additions & 2 deletions mistralrs/examples/simple/main.rs
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -4,9 +4,21 @@ use tokio::sync::mpsc::channel;
use mistralrs::{
Constraint, Device, DeviceMapMetadata, MistralRs, MistralRsBuilder, ModelDType,
NormalLoaderBuilder, NormalLoaderType, NormalRequest, NormalSpecificConfig, Request,
RequestMessage, Response, SamplingParams, SchedulerMethod, TokenSource,
RequestMessage, Response, Result, SamplingParams, SchedulerMethod, TokenSource,
};

/// Gets the best device, cpu, cuda if compiled with CUDA
pub(crate) fn best_device() -> Result<Device> {
#[cfg(not(feature = "metal"))]
{
Device::cuda_if_available(0)
}
#[cfg(feature = "metal")]
{
Device::new_metal(0)
}
}

fn setup() -> anyhow::Result<Arc<MistralRs>> {
// Select a Mistral model
let loader = NormalLoaderBuilder::new(
Expand All @@ -24,7 +36,7 @@ fn setup() -> anyhow::Result<Arc<MistralRs>> {
None,
TokenSource::CacheToken,
&ModelDType::Auto,
&Device::cuda_if_available(0)?,
&best_device()?,
false,
DeviceMapMetadata::dummy(),
None,
Expand Down
16 changes: 14 additions & 2 deletions mistralrs/examples/xlora/main.rs
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -4,9 +4,21 @@ use tokio::sync::mpsc::channel;
use mistralrs::{
Constraint, Device, DeviceMapMetadata, MistralRs, MistralRsBuilder, ModelDType,
NormalLoaderBuilder, NormalLoaderType, NormalRequest, NormalSpecificConfig, Request,
RequestMessage, Response, SamplingParams, SchedulerMethod, TokenSource,
RequestMessage, Response, Result, SamplingParams, SchedulerMethod, TokenSource,
};

/// Gets the best device, cpu, cuda if compiled with CUDA
pub(crate) fn best_device() -> Result<Device> {
#[cfg(not(feature = "metal"))]
{
Device::cuda_if_available(0)
}
#[cfg(feature = "metal")]
{
Device::new_metal(0)
}
}

fn setup() -> anyhow::Result<Arc<MistralRs>> {
// Select a Mistral model
let loader =
Expand All @@ -33,7 +45,7 @@ fn setup() -> anyhow::Result<Arc<MistralRs>> {
None,
TokenSource::CacheToken,
&ModelDType::Auto,
&Device::cuda_if_available(0)?,
&best_device()?,
false,
DeviceMapMetadata::dummy(),
None,
Expand Down
Loading