Mirrors/llama.cpp
1
0
Fork 0
mirror of https://github.com/ggerganov/llama.cpp.git synced 2025-02-05 16:10:42 +01:00
Code Issues Packages Projects Releases Wiki Activity

sync: minja (#11499)

Browse Source
This commit is contained in:
Olivier Chafik 2025-01-30 10:30:27 +00:00 committed by GitHub
parent ffd0821c57
commit 3d804dec76
No known key found for this signature in database
GPG Key ID: B5690EEEBB952194
2 changed files with 238 additions and 133 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

252
common/chat-template.hpp
View File

@ -17,17 +17,26 @@ using json = nlohmann::ordered_json;
namespace minja {
struct chat_template_caps {
bool supports_tools = false;
bool supports_tool_calls = false;
bool supports_tool_responses = false;
bool supports_system_role = false;
bool supports_parallel_tool_calls = false;
bool supports_tool_call_id = false;
// meta-llama/Llama-3.1-8B-Instruct expects arguments to be an object.
// Most other templates (and OpenAI's API) expect the arguments object to be stringified.
bool requires_object_arguments = false;
// CohereForAI/c4ai-command-r-plus simple variant
bool requires_non_null_content = false;
// MiniMaxAI/MiniMax-Text-01 special
bool requires_typed_content = false;
};
class chat_template {
public:
private:
bool supports_tools_ = true;
// Meta-Llama-3.1-8B-Instruct's template expects arguments to be an object.
// Most other templates (and OpenAI's API) expect the arguments object to be stringified.
bool requires_object_arguments_ = false;
bool requires_typed_content_ = false;
bool supports_system_role_ = true;
bool supports_parallel_tool_calls_ = false;
chat_template_caps caps_;
std::string source_;
std::string bos_token_;
std::string eos_token_;
@ -41,15 +50,16 @@ class chat_template {
{
try {
auto prompt = apply(messages, tools, add_generation_prompt, extra_context, /* adjust_inputs= */ false);
// fprintf(stderr, "Prompt: %s\n", prompt.c_str());
// fprintf(stderr, "try_raw_render: %s\n", prompt.c_str());
return prompt;
} catch (const std::exception & e) {
// fprintf(stderr, "Error: %s\n", e.what());
// fprintf(stderr, "try_raw_render error: %s\n", e.what());
return "";
}
}
public:
chat_template(const std::string & source, const std::string & bos_token, const std::string & eos_token)
: source_(source), bos_token_(bos_token), eos_token_(eos_token)
{
@ -58,69 +68,120 @@ class chat_template {
/* .lstrip_blocks = */ true,
/* .keep_trailing_newline = */ false,
});
supports_tools_ = source.find("tools") != std::string::npos;
auto renders_string_arguments =
try_raw_render({
{
{"role", "user"},
{"content", "Hey"}
},
{
{"role", "assistant"},
{"tool_calls", json::array({
{
{"id", "call_1___"},
{"type", "function"},
{"function", {
{"arguments", "{\"code\": \"print('Hello, World!')\"}"},
{"name", "ipython"},
auto contains = [](const std::string & haystack, const std::string & needle) {
return haystack.find(needle) != std::string::npos;
};
const std::string user_needle = "<User Needle>";
const std::string sys_needle = "<System Needle>";
const json dummy_str_user_msg = {{"role", "user"}, {"content", user_needle}};
const json dummy_typed_user_msg = {{"role", "user"}, {"content", json::array({{{"type", "text"}, {"text", user_needle}}})}};
caps_.requires_typed_content =
!contains(try_raw_render(json::array({dummy_str_user_msg}), {}, false), user_needle)
&& contains(try_raw_render(json::array({dummy_typed_user_msg}), {}, false), user_needle);
const auto dummy_user_msg = caps_.requires_typed_content
? dummy_typed_user_msg
: dummy_str_user_msg;
const json needle_system_msg = {
{"role", "system"},
{"content", caps_.requires_typed_content ? json::array({{{"type", "text"}, {"text", sys_needle}}}) : json(sys_needle)},
};
caps_.supports_system_role = contains(try_raw_render({needle_system_msg, dummy_user_msg,}, {}, false), sys_needle);
auto out = try_raw_render(json::array({
dummy_user_msg
}), json::array({
{
{"name", "some_tool"},
{"type", "function"},
{"function", {
{"name", "some_tool"},
{"description", "Some tool."},
{"parameters", {
{"type", "object"},
{"properties", {
{"arg", {
{"type", "string"},
{"description", "Some argument."},
}},
},
})},
}},
{"required", json::array({ "arg" })},
}},
}},
},
}), false);
caps_.supports_tools = contains(out, "some_tool");
auto make_tool_calls_msg = [&](const json & tool_calls) {
return json {
{"role", "assistant"},
{"content", nullptr},
{"tool_calls", tool_calls},
};
};
auto make_tool_call = [](const std::string & tool_name, const json & arguments) {
return json {
{"id", "call_1___"},
{"type", "function"},
{"function", {
{"arguments", arguments},
{"name", tool_name},
}},
};
};
const json dummy_args_obj {{"argument_needle", "print('Hello, World!')"}};
// Note: the arguments are rendered in both cases, but may be double-escaped, which we don't want.
out = try_raw_render(json::array({
dummy_user_msg,
make_tool_calls_msg(json::array({make_tool_call("ipython", dummy_args_obj.dump())})),
}), {}, false);
auto tool_call_renders_str_arguments = contains(out, "\"argument_needle\":") || contains(out, "'argument_needle':");
out = try_raw_render(json::array({
dummy_user_msg,
make_tool_calls_msg(json::array({make_tool_call("ipython", dummy_args_obj)})),
}), {}, false);
auto tool_call_renders_obj_arguments = contains(out, "\"argument_needle\":") || contains(out, "'argument_needle':");
caps_.supports_tool_calls = tool_call_renders_str_arguments || tool_call_renders_obj_arguments;
caps_.requires_object_arguments = !tool_call_renders_str_arguments && tool_call_renders_obj_arguments;
auto out_empty = try_raw_render(json::array({dummy_user_msg, {{"role", "assistant"}, {"content", ""}}}), {}, false);
auto out_null = try_raw_render(json::array({dummy_user_msg, {{"role", "assistant"}, {"content", nullptr}}}), {}, false);
caps_.requires_non_null_content = contains(out_empty, user_needle) && !contains(out_null, user_needle);
if (caps_.supports_tool_calls) {
auto dummy_args = caps_.requires_object_arguments ? dummy_args_obj : json(dummy_args_obj.dump());
auto tc1 = make_tool_call("test_tool1", dummy_args);
auto tc2 = make_tool_call("test_tool2", dummy_args);
auto out = try_raw_render(json::array({
dummy_user_msg,
make_tool_calls_msg(json::array({tc1, tc2})),
}), {}, false);
caps_.supports_parallel_tool_calls = contains(out, "test_tool1") && contains(out, "test_tool2");
out = try_raw_render(json::array({
dummy_user_msg,
make_tool_calls_msg(json::array({tc1})),
{
{"role", "tool"},
{"name", "test_tool1"},
{"content", "Some response!"},
{"tool_call_id", "call_911_"},
}
}, {}, false).find("{\"code\": \"print") != std::string::npos;
if (!renders_string_arguments) {
auto renders_object_arguments =
try_raw_render({
{
{"role", "user"},
{"content", "Hey"}
},
{
{"role", "assistant"},
{"tool_calls", json::array({
{
{"id", "call_1___"},
{"type", "function"},
{"function", {
{"arguments", {
{"code", "print('Hello, World!')"},
}},
{"name", "ipython"},
}},
},
})},
}
}, {}, false).find("{\"code\": \"print") != std::string::npos;
requires_object_arguments_ = renders_object_arguments;
}), {}, false);
caps_.supports_tool_responses = contains(out, "Some response!");
caps_.supports_tool_call_id = contains(out, "call_911_");
}
supports_parallel_tool_calls_ = source.find("tool_call_id") != std::string::npos;
supports_system_role_ = try_raw_render({
{{"role", "system"}, {"content", "<System Needle>"}},
{{"role", "user"}, {"content", "Hey"}}
}, {}, false).find("<System Needle>") != std::string::npos;
requires_typed_content_ = try_raw_render({{{"role", "user"}, {"content", "Hey"}}}, {}, false).find("Hey") == std::string::npos
&& try_raw_render({{{"role", "user"}, {"content", {{{"type", "text"}, {"text", "Hey"}}}}}}, {}, false).find("Hey") != std::string::npos;
}
const std::string & source() const { return source_; }
const std::string & bos_token() const { return bos_token_; }
const std::string & eos_token() const { return eos_token_; }
bool supports_tools() const { return supports_tools_; }
bool supports_parallel_tool_calls() const { return supports_parallel_tool_calls_; }
const chat_template_caps & original_caps() const { return caps_; }
std::string apply(
const nlohmann::ordered_json & messages,
@ -131,13 +192,19 @@ class chat_template {
{
json actual_messages;
// First, "fix" messages so they have a chance to be rendered correctly by the template
if (adjust_inputs && (requires_object_arguments_ || !supports_system_role_ || !supports_tools_ || requires_typed_content_)) {
auto needs_adjustments = adjust_inputs && (false
|| !caps_.supports_system_role
|| !caps_.supports_tools
|| !caps_.supports_tool_responses
|| !caps_.supports_tool_calls
|| caps_.requires_object_arguments
|| caps_.requires_typed_content
);
if (needs_adjustments) {
actual_messages = json::array();
auto add_message = [&](const json & msg) {
if (requires_typed_content_ && msg.contains("content") && !msg.at("content").is_null() && msg.at("content").is_string()) {
if (caps_.requires_typed_content && msg.contains("content") && !msg.at("content").is_null() && msg.at("content").is_string()) {
actual_messages.push_back({
{"role", msg.at("role")},
{"content", {{
@ -160,7 +227,9 @@ class chat_template {
pending_system.clear();
}
};
for (const auto & message_ : messages) {
auto needs_tools_in_system = !tools.is_null() && tools.size() > 0 && !caps_.supports_tools;
for (const auto & message_ : needs_tools_in_system ? add_system(messages, "Available tools: " + tools.dump(2)) : messages) {
auto message = message_;
if (!message.contains("role") || !message.contains("content")) {
throw std::runtime_error("message must have 'role' and 'content' fields: " + message.dump());
@ -168,16 +237,22 @@ class chat_template {
std::string role = message.at("role");
if (message.contains("tool_calls")) {
if (requires_object_arguments_ || !supports_tools_) {
if (caps_.requires_object_arguments || !caps_.supports_tool_calls) {
for (auto & tool_call : message.at("tool_calls")) {
if (tool_call["type"] == "function") {
auto & function = tool_call.at("function");
std::string arguments = function.at("arguments");
function["arguments"] = json::parse(arguments);
auto & arguments = function.at("arguments");
if (arguments.is_string()) {
try {
arguments = json::parse(arguments.get<std::string>());
} catch (const std::exception & ecvt) {
fprintf(stderr, "Failed to parse arguments: %s\n", ecvt.what());
}
}
}
}
}
if (!supports_tools_) {
if (!caps_.supports_tool_calls) {
auto content = message.at("content");
auto tool_calls = json::array();
for (const auto & tool_call : message.at("tool_calls")) {
@ -204,7 +279,7 @@ class chat_template {
message.erase("tool_calls");
}
}
if (!supports_tools_ && role == "tool") {
if (!caps_.supports_tool_responses && role == "tool") {
message["role"] = "user";
auto obj = json {
{"tool_response", {
@ -219,7 +294,7 @@ class chat_template {
message.erase("name");
}
if (!message["content"].is_null() && !supports_system_role_) {
if (!message["content"].is_null() && !caps_.supports_system_role) {
std::string content = message.at("content");
if (role == "system") {
if (!pending_system.empty()) pending_system += "\n";
@ -238,7 +313,9 @@ class chat_template {
}
add_message(message);
}
flush_sys();
if (!caps_.supports_system_role) {
flush_sys();
}
} else {
actual_messages = messages;
}
@ -261,7 +338,28 @@ class chat_template {
}
}
return template_root_->render(context);
auto ret = template_root_->render(context);
// fprintf(stderr, "actual_messages: %s\n", actual_messages.dump(2).c_str());
// fprintf(stderr, "apply: %s\n\n", ret.c_str());
return ret;
}
static nlohmann::ordered_json add_system(const nlohmann::ordered_json & messages, const std::string & system_prompt) {
json messages_with_system = messages;
if (messages_with_system.size() > 0 && messages_with_system[0].at("role") == "system") {
std::string existing_system = messages_with_system.at(0).at("content");
messages_with_system[0] = json {
{"role", "system"},
{"content", existing_system + "\n" + system_prompt},
};
} else {
messages_with_system.insert(messages_with_system.begin(), json {
{"role", "system"},
{"content", system_prompt},
});
}
return messages_with_system;
}
};

119
common/minja.hpp
View File

@ -628,7 +628,7 @@ class Context : public std::enable_shared_from_this<Context> {
if (parent_) return parent_->contains(key);
return false;
}
virtual void set(const Value & key, Value & value) {
virtual void set(const Value & key, const Value & value) {
values_.set(key, value);
}
};
@ -2648,31 +2648,34 @@ inline std::shared_ptr<Context> Context::builtins() {
return filter.call(context, actual_args);
});
};
// https://jinja.palletsprojects.com/en/3.0.x/templates/#jinja-filters.reject
globals.set("reject", Value::callable([=](const std::shared_ptr<Context> & context, ArgumentsValue & args) {
args.expectArgs("reject", {2, (std::numeric_limits<size_t>::max)()}, {0, 0});
auto & items = args.args[0];
auto filter_fn = context->get(args.args[1]);
if (filter_fn.is_null()) throw std::runtime_error("Undefined filter: " + args.args[1].dump());
auto select_or_reject = [make_filter](bool is_select) {
return Value::callable([=](const std::shared_ptr<Context> & context, ArgumentsValue & args) {
args.expectArgs(is_select ? "select" : "reject", {2, (std::numeric_limits<size_t>::max)()}, {0, 0});
auto & items = args.args[0];
auto filter_fn = context->get(args.args[1]);
if (filter_fn.is_null()) throw std::runtime_error("Undefined filter: " + args.args[1].dump());
auto filter_args = Value::array();
for (size_t i = 2, n = args.args.size(); i < n; i++) {
filter_args.push_back(args.args[i]);
}
auto filter = make_filter(filter_fn, filter_args);
auto res = Value::array();
for (size_t i = 0, n = items.size(); i < n; i++) {
auto & item = items.at(i);
ArgumentsValue filter_args;
filter_args.args.emplace_back(item);
auto pred_res = filter.call(context, filter_args);
if (!pred_res.to_bool()) {
res.push_back(item);
auto filter_args = Value::array();
for (size_t i = 2, n = args.args.size(); i < n; i++) {
filter_args.push_back(args.args[i]);
}
}
return res;
}));
auto filter = make_filter(filter_fn, filter_args);
auto res = Value::array();
for (size_t i = 0, n = items.size(); i < n; i++) {
auto & item = items.at(i);
ArgumentsValue filter_args;
filter_args.args.emplace_back(item);
auto pred_res = filter.call(context, filter_args);
if (pred_res.to_bool() == (is_select ? true : false)) {
res.push_back(item);
}
}
return res;
});
};
globals.set("select", select_or_reject(/* is_select= */ true));
globals.set("reject", select_or_reject(/* is_select= */ false));
globals.set("map", Value::callable([=](const std::shared_ptr<Context> & context, ArgumentsValue & args) {
auto res = Value::array();
if (args.args.size() == 1 &&
@ -2720,41 +2723,45 @@ inline std::shared_ptr<Context> Context::builtins() {
if (!text.empty() && text.back() == '\n') out += "\n";
return out;
}));
globals.set("selectattr", Value::callable([=](const std::shared_ptr<Context> & context, ArgumentsValue & args) {
args.expectArgs("selectattr", {2, (std::numeric_limits<size_t>::max)()}, {0, 0});
auto & items = args.args[0];
if (items.is_null())
return Value::array();
auto attr_name = args.args[1].get<std::string>();
auto select_or_reject_attr = [](bool is_select) {
return Value::callable([=](const std::shared_ptr<Context> & context, ArgumentsValue & args) {
args.expectArgs(is_select ? "selectattr" : "rejectattr", {2, (std::numeric_limits<size_t>::max)()}, {0, 0});
auto & items = args.args[0];
if (items.is_null())
return Value::array();
auto attr_name = args.args[1].get<std::string>();
bool has_test = false;
Value test_fn;
ArgumentsValue test_args {{Value()}, {}};
if (args.args.size() >= 3) {
has_test = true;
test_fn = context->get(args.args[2]);
if (test_fn.is_null()) throw std::runtime_error("Undefined test: " + args.args[2].dump());
for (size_t i = 3, n = args.args.size(); i < n; i++) {
test_args.args.emplace_back(args.args[i]);
}
test_args.kwargs = args.kwargs;
}
auto res = Value::array();
for (size_t i = 0, n = items.size(); i < n; i++) {
auto & item = items.at(i);
auto attr = item.get(attr_name);
if (has_test) {
test_args.args[0] = attr;
if (test_fn.call(context, test_args).to_bool()) {
res.push_back(item);
bool has_test = false;
Value test_fn;
ArgumentsValue test_args {{Value()}, {}};
if (args.args.size() >= 3) {
has_test = true;
test_fn = context->get(args.args[2]);
if (test_fn.is_null()) throw std::runtime_error("Undefined test: " + args.args[2].dump());
for (size_t i = 3, n = args.args.size(); i < n; i++) {
test_args.args.emplace_back(args.args[i]);
}
} else {
res.push_back(attr);
test_args.kwargs = args.kwargs;
}
}
return res;
}));
auto res = Value::array();
for (size_t i = 0, n = items.size(); i < n; i++) {
auto & item = items.at(i);
auto attr = item.get(attr_name);
if (has_test) {
test_args.args[0] = attr;
if (test_fn.call(context, test_args).to_bool() == (is_select ? true : false)) {
res.push_back(item);
}
} else {
res.push_back(attr);
}
}
return res;
});
};
globals.set("selectattr", select_or_reject_attr(/* is_select= */ true));
globals.set("rejectattr", select_or_reject_attr(/* is_select= */ false));
globals.set("range", Value::callable([=](const std::shared_ptr<Context> &, ArgumentsValue & args) {
std::vector<int64_t> startEndStep(3);
std::vector<bool> param_set(3);