mtmd: Add support for Reka Edge 2603 (#21616)

* feat: (vocab) fix stray text appended in llama_decode_text

Remove accidental concatenation of the full `text` string when
formatting UNK_BYTE hex escapes. Only the closing "]" should be appended.

* feat(mtmd): add Yasa2 vision encoder support

Add a Yasa2 (ConvNeXtV2-based) vision encoder for reka-edge:
- Register PROJECTOR_TYPE_YASA2 and tensor name definitions
- Add yasa2_block/yasa2_stage model structs
- Implement graph builder with ConvNeXt stages, GRN, adaptive pooling
- Wire into clip.cpp switch statements and mtmd.cpp init_vision
- Use mtmd_image_preprocessor_fixed_size for image preprocessing

* feat(chat): add reka-edge template handler (tools, thinking)

- Add chat-reka.cpp/h implementing PEG-based parser for reka-edge format
- Add Reka-Edge.jinja chat template
- Detect reka-edge template in try_specialized_template()
- Add LLAMA_EXAMPLE_MTMD to chat-template-file arg

* feat: add reka vlm to gguf conversion script

Converts Reka Yasa2 hf checkpoints to GGUF format:
- Text decoder: Llama-arch with tiktoken/BPE vocab
- Mmproj (--mmproj): ConvNeXt vision backbone + language_projection
- Generates 2D sincos positional embeddings for vision encoder

* test: add Reka Edge chat template and parser tests

- test-chat-template: oracle tests comparing Jinja engine output vs
  common_chat_templates_apply for text, tools, thinking, images, video
- test-chat: PEG parser tests for Reka Edge format, round-trip tests
  for image/video content parts, common path integration tests

* scripts: add Reka Edge mixed quantization helper

Q4_0 base quantization with Q8_0 override for the last 8 transformer
blocks (layers 24-31) via --tensor-type regex.

* fix: adapt chat-reka and tests to upstream API

- Use autoparser::generation_params (not templates_params)
- Add p.prefix(generation_prompt) to PEG parser
- Simplify reasoning parser to match LFM2 pattern
- Remove image/video oracle tests (unsupported by oaicompat parser;
  no other multimodal models test this path)

* fix: avoid duplicate tensor loading in yasa2 vision encoder

TN_YASA_PATCH_W and TN_PATCH_EMBD both resolve to "v.patch_embd.weight",
causing the same tensor to be loaded twice into ctx_data and overflowing
the memory pool. Reuse the tensors already loaded by the common section.

* chore: update image pre-processing settings

The reka-edge model depends on the following settings in an older
fork of llama.cpp:
1. Fixed square resize
2. BICUBIC
3. add_padding=false

In current llama.cpp, this means setting:
- image_resize_algo = RESIZE_ALGO_BICUBIC
- image_resize_pad = false

* chore: remove reka gguf conversion script

* chore: remove reka quantization script

* chore: remove unnecessary changes from PR scope

This commit removes a couple of unnecessary changes for the PR scope:
1. BPE decoder bug fix - this affects reka edge because there's a bug
in our tokenization that doesn't represent <think> tokens as special
tokens. However this isn't meant to be a thinking model so when run
with --reasoning off the edge case does not affect us

2. --chat-template-file support from llama-mtmd-cli - the focus is on
llama-server and the reka edge gguf contains the necessary metadata
to detect the chat template

3. reka edge oracle test cases - no other model has similar test cases,
so I removed it for standardization

* chore: remove unnecessary ggml_cast

This commit removes unnecessary ggml_cast after updating the
reka vlm -> gguf conversion script on hugging face.

* chore: remove redundant code

* chore: remove unnecessary ggml_cont calls

This commit removes all ggml_cont calls except the four that
precede ggml_reshape_3d/ggml_reshape_4d. Those are necessary
because ggml_reshape recomputes strides assuming contiguous
layout and asserts ggml_is_contiguous.

Other operations (ggml_mean, ggml_add, ggml_mul etc.) use
stride-based indexing and handle non-contiguous inputs
correctly and so we are ok to remove ggml_cont for those.

* chore: remove unnecessary ggml_repeat calls

This commit removes unnecessary ggml_repeat calls because the underlying
ops already broadcast automatically.

Every ggml_repeat in yasa2.cpp was expanding a smaller tensor to match
a larger one's shape before passing both to an elementwise op (ggml_add,
ggml_sub, ggml_mul, or ggml_div). This is unnecessary because all four
of these ops already support broadcasting internally.

* chore: restore ggml_cont needed for cpu operations

* refactor: locate reka chat template handler in chat.cpp

* chore: remove unnecessary warmup tokens

* chore: add code comments on image_resize_pad

* chore: remove custom reka parsing code

* chore: revert common/chat.cpp

* Uncomment debug logging for PEG input parsing

---------

Co-authored-by: Piotr Wilkin (ilintar) <piotr.wilkin@syndatis.com>

tests/test-chat.cpp

@@ -3595,6 +3595,51 @@ static void test_template_output_peg_parsers(bool detailed_debug) {
             .run();
     }
 
+    // Reka Edge
+    {
+        auto tst = peg_tester("models/templates/Reka-Edge.jinja", detailed_debug);
+        tst.test("Hello, world!\nWhat's up?")
+            .enable_thinking(false)
+            .expect(message_assist)
+            .run();
+        tst.test("I'm\nthinking</think>\n\nHello, world!\nWhat's up?")
+            .enable_thinking(true)
+            .reasoning_format(COMMON_REASONING_FORMAT_DEEPSEEK)
+            .expect(message_assist_thoughts)
+            .run();
+        tst.test("<tool_call>\n{\"name\": \"special_function\", \"arguments\": {\"arg1\": 1}}\n</tool_call>")
+            .enable_thinking(false)
+            .tools({ special_function_tool })
+            .expect(message_assist_call)
+            .run();
+        tst.test("Hello, world!\nWhat's up?\n<tool_call>\n{\"name\": \"special_function\", \"arguments\": {\"arg1\": 1}}\n</tool_call>")
+            .enable_thinking(false)
+            .tools({ special_function_tool })
+            .expect(message_assist_call_content)
+            .run();
+        tst.test("I'm\nthinking</think>\n\n<tool_call>\n{\"name\": \"special_function\", \"arguments\": {\"arg1\": 1}}\n</tool_call>")
+            .enable_thinking(true)
+            .reasoning_format(COMMON_REASONING_FORMAT_DEEPSEEK)
+            .tools({ special_function_tool })
+            .expect(message_assist_call_thoughts)
+            .run();
+        tst.test("<tool_call>\n{\"name\": \"special_function\", \"arguments\": {\"arg1\": 1}}\n</tool_call>\n<tool_call>\n{\"name\": \"special_function_with_opt\", \"arguments\": {\"arg1\": 1, \"arg2\": 2}}\n</tool_call>")
+            .enable_thinking(false)
+            .parallel_tool_calls(true)
+            .tools({ special_function_tool, special_function_tool_with_optional_param })
+            .expect_tool_calls({
+                { "special_function", R"({"arg1": 1})", {} },
+                { "special_function_with_opt", R"({"arg1": 1, "arg2": 2})", {} },
+            })
+            .run();
+        tst.test("<tool_call>\n{\"name\": \"special_function\", \"arguments\": {\"arg")
+            .enable_thinking(false)
+            .tools({ special_function_tool })
+            .is_partial(true)
+            .expect(message_assist_call_cutoff_args)
+            .run();
+    }
+
     // Apriel 1.5
     {
         auto tst = peg_tester("models/templates/unsloth-Apriel-1.5.jinja", detailed_debug);
@@ -4077,6 +4122,55 @@ static void test_template_output_peg_parsers(bool detailed_debug) {
     }
 }
 
+static void test_reka_edge_common_path() {
+    auto tmpls = read_templates("models/templates/Reka-Edge.jinja");
+
+    {
+        common_chat_templates_inputs inputs;
+        common_chat_msg system_msg;
+        system_msg.role = "system";
+        system_msg.content = "Use tools when needed.";
+
+        common_chat_msg tool_call_msg = simple_assist_msg("", "", "special_function", "{\"arg1\": 1}");
+
+        common_chat_msg tool_msg;
+        tool_msg.role = "tool";
+        tool_msg.tool_name = "special_function";
+        tool_msg.tool_call_id = "call0";
+        tool_msg.content = "Sunny";
+
+        inputs.messages = { system_msg, message_user, tool_call_msg, tool_msg, message_user };
+        inputs.tools = { special_function_tool };
+        inputs.enable_thinking = true;
+        inputs.add_generation_prompt = true;
+
+        auto params = common_chat_templates_apply(tmpls.get(), inputs);
+
+        if (params.prompt.find("<tool_response>\nSunny\n</tool_response><sep>") == std::string::npos) {
+            throw std::runtime_error("Reka Edge prompt did not render tool response history");
+        }
+        if (params.prompt.rfind("assistant: <think>\n") == std::string::npos) {
+            throw std::runtime_error("Reka Edge prompt did not render thinking generation prompt");
+        }
+    }
+
+    {
+        common_chat_templates_inputs inputs;
+        inputs.messages = {
+            message_user,
+            simple_assist_msg("The first point is")
+        };
+        inputs.add_generation_prompt = false;
+        inputs.enable_thinking = false;
+        inputs.chat_template_kwargs["continue_final_message"] = "true";
+
+        auto params = common_chat_templates_apply(tmpls.get(), inputs);
+        if (string_ends_with(params.prompt, "<sep>")) {
+            throw std::runtime_error("Reka Edge continue_final_message unexpectedly closed the assistant turn");
+        }
+    }
+}
+
 // Test the developer role to system workaround with a simple mock template
 static void test_developer_role_to_system_workaround() {
     LOG_DBG("%s\n", __func__);
@@ -4256,6 +4350,7 @@ int main(int argc, char ** argv) {
         test_msgs_oaicompat_json_conversion();
         test_tools_oaicompat_json_conversion();
         test_developer_role_to_system_workaround();
+        test_reka_edge_common_path();
         test_template_output_peg_parsers(detailed_debug);
         std::cout << "\n[chat] All tests passed!" << '\n';
     }