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

Fix mistral output repetition with F32 rope and penalty & temperature parameters #55

Merged
merged 1 commit into from
Jul 11, 2024
Merged

Fix mistral output repetition with F32 rope and penalty & temperature parameters #55

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
6 changes: 5 additions & 1 deletion README.md
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -155,7 +155,11 @@ For chat history settings, set `record_conversation` to `true` to let candle-vll

For chat streaming, the `stream` flag in chat request need to be set to `True`.

You may revise `repetition_penalty` and `temperature` flag in chat request (http post).
You may supply `penalty` and `temperature` to the model to **prevent potential repetitions**, for example:

```
cargo run --release -- --port 2000 --weight-path /home/mistral_7b/ mistral --repeat-last-n 32 --penalty 1.1 temperature 0.8
```

## Report issue
Installing `candle-vllm` is as simple as the following steps. If you have any problems, please create an
Expand Down
28 changes: 24 additions & 4 deletions src/lib.rs
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -15,6 +15,9 @@ pub enum ModelSelected {
#[arg(long)]
repeat_last_n: Option<usize>,

#[arg(long)]
temperature: Option<f32>,

#[arg(long)]
penalty: Option<f32>,
},
Expand All @@ -25,6 +28,9 @@ pub enum ModelSelected {
#[arg(long)]
repeat_last_n: Option<usize>,

#[arg(long)]
temperature: Option<f32>,

#[arg(long)]
penalty: Option<f32>,
},
Expand Down Expand Up @@ -73,6 +79,9 @@ pub enum ModelSelected {
#[arg(long)]
repeat_last_n: Option<usize>,

#[arg(long)]
temperature: Option<f32>,

#[arg(long)]
penalty: Option<f32>,
},
Expand All @@ -83,6 +92,9 @@ pub enum ModelSelected {
#[arg(long)]
repeat_last_n: Option<usize>,

#[arg(long)]
temperature: Option<f32>,

#[arg(long)]
penalty: Option<f32>,
},
Expand All @@ -93,10 +105,12 @@ impl ToString for ModelSelected {
match self {
ModelSelected::Llama {
repeat_last_n: _,
temperature: _,
penalty: _,
} => "llama".to_string(),
ModelSelected::Phi2 {
repeat_last_n: _,
temperature: _,
penalty: _,
} => "phi2".to_string(),
ModelSelected::Phi3 {
Expand All @@ -115,10 +129,12 @@ impl ToString for ModelSelected {
} => "qwen2".to_string(),
ModelSelected::Gemma {
repeat_last_n: _,
temperature: _,
penalty: _,
} => "gemma".to_string(),
ModelSelected::Mistral {
repeat_last_n: _,
temperature: _,
penalty: _,
} => "mistral".to_string(),
}
Expand All @@ -132,10 +148,11 @@ pub fn get_model_loader<'a>(
match selected_model {
ModelSelected::Llama {
repeat_last_n,
temperature,
penalty,
} => (
Box::new(DefaultLoader::new(
SpecificConfig::new(repeat_last_n, None, None, None, penalty),
SpecificConfig::new(repeat_last_n, temperature, None, None, penalty),
"llama".to_string(),
)),
if model_id.is_some() {
Expand All @@ -146,10 +163,11 @@ pub fn get_model_loader<'a>(
),
ModelSelected::Phi2 {
repeat_last_n,
temperature,
penalty,
} => (
Box::new(DefaultLoader::new(
SpecificConfig::new(repeat_last_n, None, None, None, penalty),
SpecificConfig::new(repeat_last_n, temperature, None, None, penalty),
"phi2".to_string(),
)),
if model_id.is_some() {
Expand Down Expand Up @@ -194,10 +212,11 @@ pub fn get_model_loader<'a>(
),
ModelSelected::Gemma {
repeat_last_n,
temperature,
penalty,
} => (
Box::new(DefaultLoader::new(
SpecificConfig::new(repeat_last_n, None, None, None, penalty),
SpecificConfig::new(repeat_last_n, temperature, None, None, penalty),
"gemma".to_string(),
)),
if model_id.is_some() {
Expand All @@ -208,10 +227,11 @@ pub fn get_model_loader<'a>(
),
ModelSelected::Mistral {
repeat_last_n,
temperature,
penalty,
} => (
Box::new(DefaultLoader::new(
SpecificConfig::new(repeat_last_n, None, None, None, penalty),
SpecificConfig::new(repeat_last_n, temperature, None, None, penalty),
"mistral".to_string(),
)),
if model_id.is_some() {
Expand Down
17 changes: 11 additions & 6 deletions src/openai/models/mistral.rs
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -69,9 +69,9 @@ impl RotaryEmbedding {
.map(|i| 1f32 / rope_theta.powf(i as f32 / dim as f32))
.collect();
let inv_freq_len = inv_freq.len();
let inv_freq = Tensor::from_vec(inv_freq, (1, inv_freq_len), dev)?.to_dtype(dtype)?;
let inv_freq = Tensor::from_vec(inv_freq, (1, inv_freq_len), dev)?.to_dtype(DType::F32)?;
let t = Tensor::arange(0u32, max_seq_len as u32, dev)?
.to_dtype(dtype)?
.to_dtype(DType::F32)?
.reshape((max_seq_len, 1))?;
let freqs = t.matmul(&inv_freq)?;

Expand Down Expand Up @@ -204,12 +204,17 @@ impl Attention {
let v = value_states
.reshape((b_sz, seq_len, self.num_kv_heads, self.head_dim))?
.transpose(1, 2)?;
(q.contiguous()?, k.contiguous()?, v.contiguous()?)
(q, k, v.contiguous()?)
};

let (q, k) = self
.rotary_emb
.apply_rotary_emb_qkv(&q, &k, seqlen_offset)?;
let (q, k) = self.rotary_emb.apply_rotary_emb_qkv(
&q.to_dtype(DType::F32)?,
&k.to_dtype(DType::F32)?,
seqlen_offset,
)?;

let q = q.to_dtype(v.dtype())?;
let k = k.to_dtype(v.dtype())?;

let y = self.attn.forward(
&q,
Expand Down
2 changes: 1 addition & 1 deletion src/openai/openai_server.rs
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -171,7 +171,7 @@ async fn chat_completions(
}
let sampling_params = sampling_params.unwrap();

println!("{:?}", sampling_params);
// println!("{:?}", sampling_params);

let created = get_created_time_secs();

Expand Down
16 changes: 8 additions & 8 deletions src/openai/pipelines/pipeline.rs
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -160,7 +160,7 @@ impl<'a> ModelLoader<'a> for DefaultLoader {
dtype: DType,
device: Device,
) -> Result<(Box<dyn ModulePipeline<'a>>, PipelineConfig), APIError> {
let args = self.config.clone();
let specific_args = self.config.clone();

let config = match self.name.as_str() {
"llama" => {
Expand Down Expand Up @@ -260,9 +260,9 @@ impl<'a> ModelLoader<'a> for DefaultLoader {
let pipeline_config = PipelineConfig {
max_model_len: config.max_seq_len,
default_max_tokens,
penalty: self.config.penalty.unwrap_or(1.1),
repeat_last_n: self.config.repeat_last_n.unwrap_or(16),
temperature: self.config.temperature.unwrap_or(0.),
penalty: specific_args.penalty.unwrap_or(1.1),
repeat_last_n: specific_args.repeat_last_n.unwrap_or(32),
temperature: specific_args.temperature.unwrap_or(0.),
};

println!("{:?}", pipeline_config);
Expand All @@ -272,14 +272,14 @@ impl<'a> ModelLoader<'a> for DefaultLoader {
None => tokenizer.tokenizer().token_to_id(EOS_TOKEN).unwrap(),
};

println!("{:?}", self.config);
println!("{:?}", specific_args);

let logits_processor = {
let temperature = args.temperature.unwrap_or(0.) as f64;
let temperature = specific_args.temperature.unwrap_or(0.) as f64;
let sampling = if temperature <= 0. {
Sampling::ArgMax
} else {
match (args.top_k, args.top_p) {
match (specific_args.top_k, specific_args.top_p) {
(None, None) => Sampling::All { temperature },
(Some(k), None) => Sampling::TopK { k, temperature },
(None, Some(p)) => Sampling::TopP { p, temperature },
Expand All @@ -292,7 +292,7 @@ impl<'a> ModelLoader<'a> for DefaultLoader {
Ok((
Box::new(DefaultPipeline {
model,
args,
args: specific_args,
tokenizer,
logits_processor: logits_processor,
conversation: DefaultConversation::new(
Expand Down
Loading