fix(neuron): render HF chat templates via minijinja pycompat
All checks were successful
build-prerelease / Resolve version stamps (push) Successful in 29s
CI / Format (push) Successful in 34s
CI / CUDA type-check (push) Successful in 39s
CI / Clippy (push) Successful in 2m35s
build-prerelease / Build cortex binary (push) Successful in 4m21s
build-prerelease / Build neuron-blackwell (push) Successful in 6m4s
CI / Test (push) Successful in 6m47s
CI / Build cortex SRPM (push) Has been skipped
CI / Publish cortex to COPR (push) Has been skipped
CI / Build neuron SRPM (push) Has been skipped
CI / Publish neuron to COPR (push) Has been skipped
CI / Bump version in source (push) Has been skipped
build-prerelease / Build neuron-ampere (push) Successful in 7m43s
build-prerelease / Package cortex RPM (push) Successful in 1m21s
build-prerelease / Build neuron-ada (push) Successful in 5m41s
build-prerelease / Package helexa-neuron-ada RPM (push) Successful in 3m5s
build-prerelease / Package helexa-neuron-ampere RPM (push) Successful in 3m6s
build-prerelease / Package helexa-neuron-blackwell RPM (push) Successful in 3m52s
build-prerelease / Publish to rpm.lair.cafe (unstable) (push) Successful in 1m3s
All checks were successful
build-prerelease / Resolve version stamps (push) Successful in 29s
CI / Format (push) Successful in 34s
CI / CUDA type-check (push) Successful in 39s
CI / Clippy (push) Successful in 2m35s
build-prerelease / Build cortex binary (push) Successful in 4m21s
build-prerelease / Build neuron-blackwell (push) Successful in 6m4s
CI / Test (push) Successful in 6m47s
CI / Build cortex SRPM (push) Has been skipped
CI / Publish cortex to COPR (push) Has been skipped
CI / Build neuron SRPM (push) Has been skipped
CI / Publish neuron to COPR (push) Has been skipped
CI / Bump version in source (push) Has been skipped
build-prerelease / Build neuron-ampere (push) Successful in 7m43s
build-prerelease / Package cortex RPM (push) Successful in 1m21s
build-prerelease / Build neuron-ada (push) Successful in 5m41s
build-prerelease / Package helexa-neuron-ada RPM (push) Successful in 3m5s
build-prerelease / Package helexa-neuron-ampere RPM (push) Successful in 3m6s
build-prerelease / Package helexa-neuron-blackwell RPM (push) Successful in 3m52s
build-prerelease / Publish to rpm.lair.cafe (unstable) (push) Successful in 1m3s
The Qwen3.6 chat_template.jinja (now loaded after the precedence fix) failed to render in minijinja: it uses Python str methods (content.startswith/endswith/split/rstrip/lstrip) and the raise_exception global that HF transformers patches into its Jinja env but minijinja doesn't provide. The render error tripped the text-only fallback, so image requests still produced zero <|image_pad|> tokens. Wire the standard bridge into render_chat_template: - minijinja-contrib `pycompat::unknown_method_callback` supplies the Python string/list/dict methods; - a `raise_exception` global maps to a render error (so malformed inputs — e.g. an image in a system message — surface cleanly). Add the real Qwen3.6-27B chat_template.jinja (verbatim from beast's HF cache) as a test fixture and assert it renders one <|image_pad|> for a text+image turn — the end-to-end check that would have caught this before deploy. Refs #16 / TP-vision. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
This commit is contained in:
11
Cargo.lock
generated
11
Cargo.lock
generated
@@ -2508,6 +2508,16 @@ dependencies = [
|
|||||||
"serde_json",
|
"serde_json",
|
||||||
]
|
]
|
||||||
|
|
||||||
|
[[package]]
|
||||||
|
name = "minijinja-contrib"
|
||||||
|
version = "2.20.0"
|
||||||
|
source = "registry+https://github.com/rust-lang/crates.io-index"
|
||||||
|
checksum = "99df5123c54391e2a228014c1dbbd85a3dab08a25e776c810526f2f47542b3de"
|
||||||
|
dependencies = [
|
||||||
|
"minijinja",
|
||||||
|
"serde",
|
||||||
|
]
|
||||||
|
|
||||||
[[package]]
|
[[package]]
|
||||||
name = "minimal-lexical"
|
name = "minimal-lexical"
|
||||||
version = "0.2.1"
|
version = "0.2.1"
|
||||||
@@ -2605,6 +2615,7 @@ dependencies = [
|
|||||||
"hf-hub",
|
"hf-hub",
|
||||||
"image",
|
"image",
|
||||||
"minijinja",
|
"minijinja",
|
||||||
|
"minijinja-contrib",
|
||||||
"reqwest",
|
"reqwest",
|
||||||
"safetensors 0.7.0",
|
"safetensors 0.7.0",
|
||||||
"serde",
|
"serde",
|
||||||
|
|||||||
@@ -76,15 +76,19 @@ cudarc = { version = "0.19", optional = true, default-features = false, features
|
|||||||
half = { version = "2.5", optional = true }
|
half = { version = "2.5", optional = true }
|
||||||
tokenizers = { version = "0.22", default-features = false, features = ["onig"] }
|
tokenizers = { version = "0.22", default-features = false, features = ["onig"] }
|
||||||
hf-hub = { version = "0.4", features = ["tokio"] }
|
hf-hub = { version = "0.4", features = ["tokio"] }
|
||||||
# Jinja-compatible template renderer for the model's
|
# Jinja-compatible template renderer for the model's chat template
|
||||||
# `tokenizer_config.json::chat_template`. Hugging Face's chat
|
# (standalone `chat_template.jinja` or `tokenizer_config.json::chat_template`).
|
||||||
# templates use a strict subset of Jinja2 that minijinja supports
|
# Hugging Face's chat templates lean on Python string semantics; we
|
||||||
# out of the box. ~80KB compiled; pure Rust, no async surface.
|
# bridge them with `minijinja-contrib`'s `pycompat` callback (str
|
||||||
# Features: `builtins` for the `is defined` / `default` filters HF
|
# methods like `startswith`/`split`/`strip`) plus a `raise_exception`
|
||||||
# templates use; `json` for `tojson` (some Qwen3 templates emit
|
# global. Features: `builtins` for `is defined` / `default`; `json`
|
||||||
# tool definitions via tojson); `serde` so we can hand it a
|
# for `tojson`; `serde` so we can hand it a serde_json::Value context.
|
||||||
# serde_json::Value as the context.
|
|
||||||
minijinja = { version = "2", features = ["builtins", "json", "serde"] }
|
minijinja = { version = "2", features = ["builtins", "json", "serde"] }
|
||||||
|
# Python-compatibility shim: the Qwen3-VL / Qwen3.6 template uses
|
||||||
|
# `content.startswith(...)`, `.endswith(...)`, `.split(...)`,
|
||||||
|
# `.rstrip(...)`, `.lstrip(...)` — Python str methods minijinja doesn't
|
||||||
|
# implement natively. `pycompat::unknown_method_callback` supplies them.
|
||||||
|
minijinja-contrib = { version = "2", features = ["pycompat"] }
|
||||||
# Direct dep on `safetensors` (re-exported by candle but its `TensorView`
|
# Direct dep on `safetensors` (re-exported by candle but its `TensorView`
|
||||||
# / `slice::IndexOp` types are public-but-not-re-exported). Used by the
|
# / `slice::IndexOp` types are public-but-not-re-exported). Used by the
|
||||||
# tp `fused_load` module to read per-rank slices of fused QKV tensors
|
# tp `fused_load` module to read per-rank slices of fused QKV tensors
|
||||||
|
|||||||
@@ -43,7 +43,7 @@
|
|||||||
|
|
||||||
use anyhow::{Context, Result};
|
use anyhow::{Context, Result};
|
||||||
use cortex_core::openai::{ChatMessage, MessageContent};
|
use cortex_core::openai::{ChatMessage, MessageContent};
|
||||||
use minijinja::Environment;
|
use minijinja::{Environment, Error as MjError, ErrorKind as MjErrorKind, Value as MjValue};
|
||||||
use serde_json::Value;
|
use serde_json::Value;
|
||||||
use std::path::Path;
|
use std::path::Path;
|
||||||
|
|
||||||
@@ -191,6 +191,25 @@ pub fn render_chat_template(
|
|||||||
kwargs: &Value,
|
kwargs: &Value,
|
||||||
) -> Result<String> {
|
) -> Result<String> {
|
||||||
let mut env = Environment::new();
|
let mut env = Environment::new();
|
||||||
|
|
||||||
|
// HF chat templates are authored against Python's Jinja2 with its
|
||||||
|
// string semantics. Bridge the two so real model templates render:
|
||||||
|
//
|
||||||
|
// - `pycompat::unknown_method_callback` supplies Python str/list/dict
|
||||||
|
// methods minijinja lacks natively (`startswith`, `endswith`,
|
||||||
|
// `split`, `rstrip`, `lstrip`, …) — the Qwen3.6 template uses
|
||||||
|
// several in its think-block and tool-response handling.
|
||||||
|
// - `raise_exception` is the global HF templates call to reject
|
||||||
|
// malformed inputs (e.g. an image in a system message). Map it to
|
||||||
|
// a render error so the caller falls back / surfaces it.
|
||||||
|
env.set_unknown_method_callback(minijinja_contrib::pycompat::unknown_method_callback);
|
||||||
|
env.add_function(
|
||||||
|
"raise_exception",
|
||||||
|
|msg: String| -> Result<MjValue, MjError> {
|
||||||
|
Err(MjError::new(MjErrorKind::InvalidOperation, msg))
|
||||||
|
},
|
||||||
|
);
|
||||||
|
|
||||||
// Compile the template against a fixed name so error messages
|
// Compile the template against a fixed name so error messages
|
||||||
// surface "chat_template" rather than `<template>`.
|
// surface "chat_template" rather than `<template>`.
|
||||||
env.add_template("chat_template", template)
|
env.add_template("chat_template", template)
|
||||||
@@ -334,6 +353,33 @@ mod tests {
|
|||||||
assert_eq!(got.as_deref(), Some("FROM_CONFIG"));
|
assert_eq!(got.as_deref(), Some("FROM_CONFIG"));
|
||||||
}
|
}
|
||||||
|
|
||||||
|
/// The *actual* Qwen3.6-27B `chat_template.jinja` (verbatim from
|
||||||
|
/// beast's HF cache) must render in minijinja and emit exactly one
|
||||||
|
/// `<|image_pad|>` for a text+image user turn. This is the real
|
||||||
|
/// end-to-end check the unit tests above only approximate — it
|
||||||
|
/// catches any minijinja incompatibility (namespace, macros,
|
||||||
|
/// reverse slice, string methods) before it reaches production.
|
||||||
|
#[test]
|
||||||
|
fn real_qwen3_6_template_renders_one_image_pad() {
|
||||||
|
let template = include_str!("testdata/qwen3_6_chat_template.jinja");
|
||||||
|
let messages = vec![ChatMessage {
|
||||||
|
role: "user".into(),
|
||||||
|
content: MessageContent::Parts(vec![
|
||||||
|
json!({"type": "text", "text": "what is this?"}),
|
||||||
|
json!({"type": "image_url", "image_url": {"url": "data:image/png;base64,AAA="}}),
|
||||||
|
]),
|
||||||
|
extra: Value::Object(Default::default()),
|
||||||
|
}];
|
||||||
|
let out = render_chat_template(template, &messages, &Value::Null, &Value::Null)
|
||||||
|
.expect("real Qwen3.6 template should render in minijinja");
|
||||||
|
let pads = out.matches("<|image_pad|>").count();
|
||||||
|
assert_eq!(
|
||||||
|
pads, 1,
|
||||||
|
"expected exactly one <|image_pad|>; rendered:\n{out}"
|
||||||
|
);
|
||||||
|
assert!(out.contains("<|vision_start|>") && out.contains("<|vision_end|>"));
|
||||||
|
}
|
||||||
|
|
||||||
fn user_msg(text: &str) -> ChatMessage {
|
fn user_msg(text: &str) -> ChatMessage {
|
||||||
ChatMessage {
|
ChatMessage {
|
||||||
role: "user".into(),
|
role: "user".into(),
|
||||||
|
|||||||
154
crates/neuron/src/harness/testdata/qwen3_6_chat_template.jinja
vendored
Normal file
154
crates/neuron/src/harness/testdata/qwen3_6_chat_template.jinja
vendored
Normal file
@@ -0,0 +1,154 @@
|
|||||||
|
{%- set image_count = namespace(value=0) %}
|
||||||
|
{%- set video_count = namespace(value=0) %}
|
||||||
|
{%- macro render_content(content, do_vision_count, is_system_content=false) %}
|
||||||
|
{%- if content is string %}
|
||||||
|
{{- content }}
|
||||||
|
{%- elif content is iterable and content is not mapping %}
|
||||||
|
{%- for item in content %}
|
||||||
|
{%- if 'image' in item or 'image_url' in item or item.type == 'image' %}
|
||||||
|
{%- if is_system_content %}
|
||||||
|
{{- raise_exception('System message cannot contain images.') }}
|
||||||
|
{%- endif %}
|
||||||
|
{%- if do_vision_count %}
|
||||||
|
{%- set image_count.value = image_count.value + 1 %}
|
||||||
|
{%- endif %}
|
||||||
|
{%- if add_vision_id %}
|
||||||
|
{{- 'Picture ' ~ image_count.value ~ ': ' }}
|
||||||
|
{%- endif %}
|
||||||
|
{{- '<|vision_start|><|image_pad|><|vision_end|>' }}
|
||||||
|
{%- elif 'video' in item or item.type == 'video' %}
|
||||||
|
{%- if is_system_content %}
|
||||||
|
{{- raise_exception('System message cannot contain videos.') }}
|
||||||
|
{%- endif %}
|
||||||
|
{%- if do_vision_count %}
|
||||||
|
{%- set video_count.value = video_count.value + 1 %}
|
||||||
|
{%- endif %}
|
||||||
|
{%- if add_vision_id %}
|
||||||
|
{{- 'Video ' ~ video_count.value ~ ': ' }}
|
||||||
|
{%- endif %}
|
||||||
|
{{- '<|vision_start|><|video_pad|><|vision_end|>' }}
|
||||||
|
{%- elif 'text' in item %}
|
||||||
|
{{- item.text }}
|
||||||
|
{%- else %}
|
||||||
|
{{- raise_exception('Unexpected item type in content.') }}
|
||||||
|
{%- endif %}
|
||||||
|
{%- endfor %}
|
||||||
|
{%- elif content is none or content is undefined %}
|
||||||
|
{{- '' }}
|
||||||
|
{%- else %}
|
||||||
|
{{- raise_exception('Unexpected content type.') }}
|
||||||
|
{%- endif %}
|
||||||
|
{%- endmacro %}
|
||||||
|
{%- if not messages %}
|
||||||
|
{{- raise_exception('No messages provided.') }}
|
||||||
|
{%- endif %}
|
||||||
|
{%- if tools and tools is iterable and tools is not mapping %}
|
||||||
|
{{- '<|im_start|>system\n' }}
|
||||||
|
{{- "# Tools\n\nYou have access to the following functions:\n\n<tools>" }}
|
||||||
|
{%- for tool in tools %}
|
||||||
|
{{- "\n" }}
|
||||||
|
{{- tool | tojson }}
|
||||||
|
{%- endfor %}
|
||||||
|
{{- "\n</tools>" }}
|
||||||
|
{{- '\n\nIf you choose to call a function ONLY reply in the following format with NO suffix:\n\n<tool_call>\n<function=example_function_name>\n<parameter=example_parameter_1>\nvalue_1\n</parameter>\n<parameter=example_parameter_2>\nThis is the value for the second parameter\nthat can span\nmultiple lines\n</parameter>\n</function>\n</tool_call>\n\n<IMPORTANT>\nReminder:\n- Function calls MUST follow the specified format: an inner <function=...></function> block must be nested within <tool_call></tool_call> XML tags\n- Required parameters MUST be specified\n- You may provide optional reasoning for your function call in natural language BEFORE the function call, but NOT after\n- If there is no function call available, answer the question like normal with your current knowledge and do not tell the user about function calls\n</IMPORTANT>' }}
|
||||||
|
{%- if messages[0].role == 'system' %}
|
||||||
|
{%- set content = render_content(messages[0].content, false, true)|trim %}
|
||||||
|
{%- if content %}
|
||||||
|
{{- '\n\n' + content }}
|
||||||
|
{%- endif %}
|
||||||
|
{%- endif %}
|
||||||
|
{{- '<|im_end|>\n' }}
|
||||||
|
{%- else %}
|
||||||
|
{%- if messages[0].role == 'system' %}
|
||||||
|
{%- set content = render_content(messages[0].content, false, true)|trim %}
|
||||||
|
{{- '<|im_start|>system\n' + content + '<|im_end|>\n' }}
|
||||||
|
{%- endif %}
|
||||||
|
{%- endif %}
|
||||||
|
{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}
|
||||||
|
{%- for message in messages[::-1] %}
|
||||||
|
{%- set index = (messages|length - 1) - loop.index0 %}
|
||||||
|
{%- if ns.multi_step_tool and message.role == "user" %}
|
||||||
|
{%- set content = render_content(message.content, false)|trim %}
|
||||||
|
{%- if not(content.startswith('<tool_response>') and content.endswith('</tool_response>')) %}
|
||||||
|
{%- set ns.multi_step_tool = false %}
|
||||||
|
{%- set ns.last_query_index = index %}
|
||||||
|
{%- endif %}
|
||||||
|
{%- endif %}
|
||||||
|
{%- endfor %}
|
||||||
|
{%- if ns.multi_step_tool %}
|
||||||
|
{{- raise_exception('No user query found in messages.') }}
|
||||||
|
{%- endif %}
|
||||||
|
{%- for message in messages %}
|
||||||
|
{%- set content = render_content(message.content, true)|trim %}
|
||||||
|
{%- if message.role == "system" %}
|
||||||
|
{%- if not loop.first %}
|
||||||
|
{{- raise_exception('System message must be at the beginning.') }}
|
||||||
|
{%- endif %}
|
||||||
|
{%- elif message.role == "user" %}
|
||||||
|
{{- '<|im_start|>' + message.role + '\n' + content + '<|im_end|>' + '\n' }}
|
||||||
|
{%- elif message.role == "assistant" %}
|
||||||
|
{%- set reasoning_content = '' %}
|
||||||
|
{%- if message.reasoning_content is string %}
|
||||||
|
{%- set reasoning_content = message.reasoning_content %}
|
||||||
|
{%- else %}
|
||||||
|
{%- if '</think>' in content %}
|
||||||
|
{%- set reasoning_content = content.split('</think>')[0].rstrip('\n').split('<think>')[-1].lstrip('\n') %}
|
||||||
|
{%- set content = content.split('</think>')[-1].lstrip('\n') %}
|
||||||
|
{%- endif %}
|
||||||
|
{%- endif %}
|
||||||
|
{%- set reasoning_content = reasoning_content|trim %}
|
||||||
|
{%- if (preserve_thinking is defined and preserve_thinking is true) or (loop.index0 > ns.last_query_index) %}
|
||||||
|
{{- '<|im_start|>' + message.role + '\n<think>\n' + reasoning_content + '\n</think>\n\n' + content }}
|
||||||
|
{%- else %}
|
||||||
|
{{- '<|im_start|>' + message.role + '\n' + content }}
|
||||||
|
{%- endif %}
|
||||||
|
{%- if message.tool_calls and message.tool_calls is iterable and message.tool_calls is not mapping %}
|
||||||
|
{%- for tool_call in message.tool_calls %}
|
||||||
|
{%- if tool_call.function is defined %}
|
||||||
|
{%- set tool_call = tool_call.function %}
|
||||||
|
{%- endif %}
|
||||||
|
{%- if loop.first %}
|
||||||
|
{%- if content|trim %}
|
||||||
|
{{- '\n\n<tool_call>\n<function=' + tool_call.name + '>\n' }}
|
||||||
|
{%- else %}
|
||||||
|
{{- '<tool_call>\n<function=' + tool_call.name + '>\n' }}
|
||||||
|
{%- endif %}
|
||||||
|
{%- else %}
|
||||||
|
{{- '\n<tool_call>\n<function=' + tool_call.name + '>\n' }}
|
||||||
|
{%- endif %}
|
||||||
|
{%- if tool_call.arguments is defined %}
|
||||||
|
{%- for args_name, args_value in tool_call.arguments|items %}
|
||||||
|
{{- '<parameter=' + args_name + '>\n' }}
|
||||||
|
{%- set args_value = args_value | string if args_value is string else args_value | tojson | safe %}
|
||||||
|
{{- args_value }}
|
||||||
|
{{- '\n</parameter>\n' }}
|
||||||
|
{%- endfor %}
|
||||||
|
{%- endif %}
|
||||||
|
{{- '</function>\n</tool_call>' }}
|
||||||
|
{%- endfor %}
|
||||||
|
{%- endif %}
|
||||||
|
{{- '<|im_end|>\n' }}
|
||||||
|
{%- elif message.role == "tool" %}
|
||||||
|
{%- if loop.previtem and loop.previtem.role != "tool" %}
|
||||||
|
{{- '<|im_start|>user' }}
|
||||||
|
{%- endif %}
|
||||||
|
{{- '\n<tool_response>\n' }}
|
||||||
|
{{- content }}
|
||||||
|
{{- '\n</tool_response>' }}
|
||||||
|
{%- if not loop.last and loop.nextitem.role != "tool" %}
|
||||||
|
{{- '<|im_end|>\n' }}
|
||||||
|
{%- elif loop.last %}
|
||||||
|
{{- '<|im_end|>\n' }}
|
||||||
|
{%- endif %}
|
||||||
|
{%- else %}
|
||||||
|
{{- raise_exception('Unexpected message role.') }}
|
||||||
|
{%- endif %}
|
||||||
|
{%- endfor %}
|
||||||
|
{%- if add_generation_prompt %}
|
||||||
|
{{- '<|im_start|>assistant\n' }}
|
||||||
|
{%- if enable_thinking is defined and enable_thinking is false %}
|
||||||
|
{{- '<think>\n\n</think>\n\n' }}
|
||||||
|
{%- else %}
|
||||||
|
{{- '<think>\n' }}
|
||||||
|
{%- endif %}
|
||||||
|
{%- endif %}
|
||||||
Reference in New Issue
Block a user