add chat template and tokens

chat_template.jinja

@@ -0,0 +1,383 @@
+{%- set system_start = '<|im_start|>system\n' -%}
+{%- set user_start = '<|im_start|>user\n' -%}
+{%- set assistant_start = '<|im_start|>assistant\n' -%}
+{%- set system_end = '<|im_end|>\n' -%}
+{%- set user_end = '<|im_end|>\n' -%}
+{%- set assistant_end = '<|im_end|>\n' -%}
+{%- set tool_defs_start = '<|tools_start|>\n' -%}
+{%- set tool_defs_end = '\n<|tools_end|>' -%}
+{%- set tool_call_start = '\n<|tool_start|>\n' -%}
+{%- set tool_call_end = '\n<|tool_end|>' -%}
+{%- set tool_result_start = '<|im_start|>tool\n' -%}
+{%- set tool_result_sep = '\n' -%}
+{%- set tool_result_end = '\n<|im_end|>\n' -%}
+{%- set reasoning_start = '<|thinking|>\n' -%}
+{%- set reasoning_end = '\n<|thinking_end|>\n' -%}
+{%- set default_system_prompt = "" -%}
+{%- set tool_defs_preamble = "You are provided with function signatures within " + tool_defs_start.strip() + tool_defs_end.strip() + "\n" -%}
+{%- set tool_defs_postamble = "\nFor each function call, call them within " + tool_call_start.strip() + tool_call_end.strip() + "." -%}
+{%- set tool_defs_style = "python" -%}
+{%- set tool_call_style = "python" -%}
+
+{%- set content_ns = namespace(content='') %}
+
+{%- if messages[0].role == 'system' %}
+    {%- if messages[0].content is string %}
+        {%- set content_ns.content = messages[0].content %}
+    {%- elif messages[0].content is iterable %}
+        {%- set content_ns.content = '' %}
+        {%- for part in messages[0].content %}
+            {%- set content_ns.content = content_ns.content + part.text %}
+        {%- endfor %}
+    {%- else %}
+        {%- set content_ns.content = '' %}
+    {%- endif %}
+    {%- set messages = messages[1:] %}
+{%- else %}
+    {%- set content_ns.content = default_system_prompt %}
+{%- endif %}
+
+{%- macro json_to_python_type(json_spec) %}
+    {%- set basic_type_map = {
+        "string": "str",
+        "number": "float",
+        "integer": "int",
+        "boolean": "bool"
+    } %}
+
+    {%- if basic_type_map[json_spec.type] is defined %}
+        {{- basic_type_map[json_spec.type] }}
+    {%- elif json_spec.type == "array" %}
+        {{- "List[" +  json_to_python_type(json_spec.items) + "]"}}
+    {%- elif json_spec.type == "object" %}
+        {{- "Dict[str, " + json_to_python_type(json_spec.additionalProperties) + ']'}}
+    {%- elif json_spec.type is iterable %}
+        {{- "Union[" }}
+        {%- for t in json_spec.type %}
+            {{- json_to_python_type({"type": t}) }}
+            {%- if not loop.last %}
+                {{- "," }} 
+            {%- endif %}
+        {%- endfor %}
+        {{- "]" }}
+    {%- else %}
+        {{- "Any" }}
+    {%- endif %}
+{%- endmacro %}
+
+{%- macro json_to_python_value(json_value) %}
+    {%- if json_value is string %}
+        {{- json_value|tojson }}
+    {%- elif json_value is integer or json_value is float %}
+        {{- json_value|tojson }}
+    {%- elif json_value is boolean %}
+        {%- if json_value %}
+            {{- "True" }}
+        {%- else %}
+            {{- "False" }}
+        {%- endif %}
+    {%- elif json_value is iterable %}
+        {{- "[" }}
+        {%- for item in json_value %}
+            {{- json_to_python_value(item) }}
+            {%- if not loop.last %}
+                {{- ',' }}
+            {%- endif %}
+        {%- endfor %}
+        {{- "]" }}
+    {%- else %}
+        {{- "None" }}
+    {%- endif %}
+{%- endmacro%}
+
+{%- macro python_tools_parser(tools) %}
+    {%- for tool in tools %}
+        {%- if loop.index0 != 0 %}
+            {{- '\n\n'}}
+        {%- endif %}
+        {%- if tool.function is defined %}
+            {%- set tool = tool.function %}
+        {%- endif %}
+        {{-'def ' + tool.name + '('}}
+        {%- for param_name, param_fields in tool.parameters.properties|items %}
+            {%- if loop.index0 != 0 %}
+                {{- ', '}}
+            {%- endif %}
+            {{-param_name + ": "}} 
+            {%- if not param_name in tool.parameters.required %}
+                {{-'Optional[' + json_to_python_type(param_fields) + '] = None'}}
+            {%- else %}
+                {{- json_to_python_type(param_fields) }}
+            {%- endif %}
+        {%- endfor %}
+        {{- ') -> str:\n    """'}}
+        {{- tool.description }}
+        {%- if tool.parameters.properties|length != 0 %}
+            {{- '\n\n    Args:\n        '}}
+            {%- for param_name, param_fields in tool.parameters.properties|items %}
+                {%- if loop.index0 != 0 %}
+                    {{- '\n        ' }}
+                {%- endif %}
+                {{- param_name + ' ('}}
+                {%- if not param_name in tool.parameters.required %}
+                    {{-'Optional[' + json_to_python_type(param_fields) + ']'}}
+                {%- else %}
+                    {{- json_to_python_type(param_fields) }}
+                {%- endif %}
+                {{- ')' }}
+                {%- if param_fields.description %}
+                    {{- ': ' + param_fields.description }}
+                {%- endif%}
+            {%- endfor %}
+        {%- endif %}
+        {{- '\n    """\n    pass' }}
+    {%- endfor %}
+{%- endmacro %}
+
+{%- macro python_tool_call_parser(tool_call) %}
+    {{- tool_call.name + '(' }}
+    {%- if tool_call.arguments is string %}
+        {{- raise_exception("Python tool call parser doesn't support arguments as a json string") }}
+    {%- endif %}
+    {%- for param_name, param_field in tool_call.arguments|items %}
+        {{- param_name + '= '}}
+        {{- json_to_python_value(param_field) }}
+        {%- if not loop.last %}
+            {{- ', '}}
+        {%- endif %}
+    {%- endfor %}
+    {{- ')' }}
+{%- endmacro %}
+
+{%- macro json_tools_parser(tools) %}
+{%- for tool in tools %}
+    {%- if not loop.first %}
+        {{- "\n" }}
+    {%- endif %}
+    {{- tool | tojson }}
+{%- endfor %}
+{%- endmacro %}
+
+{%- macro json_tool_call_parser(tool_call) %}
+    {{-  '{"name": "' }}
+    {{- tool_call.name }}
+    {{- '", "arguments": ' }}
+    {%- if tool_call.arguments is string %}
+        {{- tool_call.arguments }}
+    {%- else %}
+        {{- tool_call.arguments | tojson }}
+    {%- endif %}
+    {{- '}' }}
+{%- endmacro %}
+
+{%- if tools %}
+    {{- system_start }}
+    {%- if content_ns.content | length > 0 %}
+        {{- content_ns.content + '\n\n' }}
+    {%- endif %}
+    {{- tool_defs_preamble + tool_defs_start }}
+    {%- if tool_defs_style == "json" %}
+        {{- json_tools_parser(tools) }}
+    {%- elif tool_defs_style == "python" %}
+        {{- python_tools_parser(tools) }}
+    {%- else %}
+        {{- raise_exception("Invalid tool definition style (" + tool_defs_style + "), must be 'json' or 'python'") }}
+    {%- endif %}
+    {{- tool_defs_end + tool_defs_postamble }}
+    {%- if (messages|length == 0) and not add_generation_prompt %}
+        {{- system_end.rstrip('\n') }}
+    {%- else %}
+        {{- system_end }}
+    {%- endif %}
+{%- else %}
+    {%- if content_ns.content | length > 0 %}
+        {{- system_start + content_ns.content }}
+        {%- if (messages|length == 0) and not add_generation_prompt %}
+            {{- system_end.rstrip('\n') }}
+        {%- else %}
+            {{- system_end }}
+        {%- endif %}
+    {%- 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" and message.content is string and not(message.content.startswith(tool_result_start) and message.content.endswith(tool_result_end)) %}
+        {%- set ns.multi_step_tool = false %}
+        {%- set ns.last_query_index = index %}
+    {%- endif %}
+{%- endfor %}
+{%- for message in messages %}
+
+    {%- if message.content is string %}
+        {%- set content_ns.content = message.content %}
+    {%- elif message.content is iterable %}
+        {%- set content_ns.content = '' %}
+        {%- for part in message.content %}
+            {%- set content_ns.content = content_ns.content + part.text %}
+        {%- endfor %}
+    {%- else %}
+        {%- set content_ns.content = '' %}
+    {%- endif %}
+
+    {%- if message.role == "system" %}
+        {{- system_start }}
+
+        {%- if loop.last and not add_generation_prompt %}
+            {%- set end_marker = system_end.rstrip('\n') %}
+        {%- else %}
+            {%- set end_marker = system_end %}
+        {%- endif %}
+
+        {%- if message.content is string %}
+            {{- content_ns.content + end_marker }}
+        {%- else %}
+            {%- for item in message.content %}
+                {%- if item.mask is defined and not item.mask %}
+                    {%- generation %}
+                    {{- item.text }}
+                    {%- if loop.last %}
+                        {{- end_marker }}
+                    {%- endif %}
+                    {%- endgeneration %}
+                {%- else %}
+                    {{- item.text }}
+                    {%- if loop.last %}
+                        {{- end_marker }}
+                    {%- endif %}
+                {%- endif %}
+            {%- endfor %}
+        {%- endif %}
+
+
+    {%- elif message.role == "user" %}
+        {{- user_start }}
+
+        {%- if loop.last and not add_generation_prompt %}
+            {%- set end_marker = user_end.rstrip('\n') %}
+        {%- else %}
+            {%- set end_marker = user_end %}
+        {%- endif %}
+
+        {%- if message.content is string %}
+            {{- content_ns.content + end_marker }}
+        {%- else %}
+            {%- for item in message.content %}
+                {%- if item.mask is defined and not item.mask %}
+                    {%- generation %}
+                    {{- item.text }}
+                    {%- if loop.last %}
+                        {{- end_marker }}
+                    {%- endif %}
+                    {%- endgeneration %}
+                {%- else %}
+                    {{- item.text }}
+                    {%- if loop.last %}
+                        {{- end_marker }}
+                    {%- endif %}
+                {%- endif %}
+            {%- endfor %}
+        {%- endif %}
+
+
+    {%- elif message.role == "assistant" %}
+        {%- set reasoning_content = none %}
+        {%- if message.reasoning_content is string %}
+            {%- set reasoning_content = message.reasoning_content %}
+        {%- else %}
+            {%- if reasoning_start in content_ns.content %}
+                {%- set reasoning_content = content_ns.content.split(reasoning_end)[0].rstrip('\n').split(reasoning_start)[-1].lstrip('\n') %}
+                {%- set content = content_ns.content.split(reasoning_end)[-1].lstrip('\n') %}
+            {%- endif %}
+        {%- endif %}
+
+        {{- assistant_start }}
+
+        {%- if loop.index0 > ns.last_query_index %}
+            {%- if loop.last and reasoning_content %}
+                {%- set reasoning_end_marker = reasoning_end %}
+                {%- if content_ns.content is string %}
+                    {%- set content_ns.content = content_ns.content.lstrip('\n') %}
+                    {%- if content_ns.content == '' %}
+                        {%- set reasoning_end_marker = reasoning_end.rstrip('\n') %}
+                    {%- endif %}
+                {%- endif %}
+                {%- generation %}
+                {{- reasoning_start + reasoning_content.strip('\n') + reasoning_end_marker }}
+                {%- endgeneration %}
+            {%- endif %}
+        {%- endif %}
+
+        {%- if message.content is string %}
+            {%- generation %}
+            {{- content_ns.content }}
+            {%- endgeneration %}
+        {%- else %}
+            {%- for item in message.content %}
+                {%- if not item.mask is defined or not item.mask %}
+                    {%- generation %}
+                    {{- item.text }}
+                    {%- endgeneration %}
+                {%- else %}
+                    {{- item.text }}
+                {%- endif %}
+            {%- endfor %}
+        {%- endif %}
+        {%- generation %}
+        {%- if message.tool_calls %}
+            {%- for tool_call in message.tool_calls %}
+                {%- if tool_call.function %}
+                    {%- set tool_call = tool_call.function %}
+                {%- endif %}
+                {%- if (loop.first and content_ns.content) or (not loop.first) %}
+                    {{- tool_call_start }}
+                {%- else %}
+                    {{- tool_call_start.lstrip('\n') }}
+                {%- endif %}
+                {%- if tool_call_style == 'json' %}
+                    {{- json_tool_call_parser(tool_call) }}
+                {%- elif tool_call_style == 'python' %}
+                    {{- python_tool_call_parser(tool_call) }}
+                {%- endif %}
+                {{- tool_call_end }}
+            {%- endfor %}
+        {%- endif %}
+        {%- endgeneration %}
+        {%- if not continue_final_message %}
+            {%- if loop.last %}
+                {%- set end_marker = assistant_end.rstrip('\n') %}
+            {%- else %}
+                {%- set end_marker = assistant_end %}
+            {%- endif %}
+        {%- endif %}            
+        {%- if message.content is string or (message.content is defined and (not message.content[message.content|length - 1].mask is defined or not message.content[message.content|length - 1].mask)) or (message.tool_calls) %}
+            {%- generation%}
+            {{- end_marker }}
+            {%- endgeneration %}
+        {%- else %}
+            {{- end_marker }}
+        {%- endif %}
+
+    {%- elif message.role == "tool" %}
+        {%- if loop.first or (messages[loop.index0 - 1].role != "tool") %}
+            {{- tool_result_start }}
+        {%- else %}
+            {{- tool_result_sep }}
+        {%- endif %}
+        {{- content_ns.content }}
+        {%- if loop.last %}
+            {{- tool_result_end.rstrip('\n') }}
+        {%- elif (messages[loop.index0 + 1].role != "tool") %}
+            {{- tool_result_end.rstrip('\n') }}
+        {%- endif %}
+    {%- endif %}
+{%- endfor %}
+{%- if add_generation_prompt %}
+    {{- assistant_start }}
+    {%- if enable_thinking is defined and enable_thinking is true %}
+        {%- generation %}
+        {{- reasoning_start }}
+        {%- endgeneration %}
+    {%- endif %}
+{%- endif -%}

config.json

@@ -1,11 +1,11 @@
 {
-  "_name_or_path": "unsloth/Mistral-7B-v0.3",
+  "_name_or_path": "NyxKrage/Mistral-7B-v0.3",
   "architectures": [
     "MistralForCausalLM"
   ],
   "attention_dropout": 0.0,
   "bos_token_id": 1,
-  "eos_token_id": 2,
+  "eos_token_id": 11,
   "head_dim": 128,
   "hidden_act": "silu",
   "hidden_size": 4096,

generation_config.json

@@ -1,7 +1,7 @@
 {
   "_from_model_config": true,
   "bos_token_id": 1,
-  "eos_token_id": 2,
+  "eos_token_id": 11,
   "max_length": 32768,
   "pad_token_id": 770,
   "transformers_version": "4.44.2"

special_tokens_map.json

@@ -7,7 +7,7 @@
     "single_word": false
   },
   "eos_token": {
-    "content": "</s>",
+    "content": "<|im_end|>",
     "lstrip": false,
     "normalized": false,
     "rstrip": false,

tokenizer.json

@@ -95,7 +95,7 @@
     },
     {
       "id": 10,
-      "content": "[control_8]",
+      "content": "<|im_start|>",
       "single_word": false,
       "lstrip": false,
       "rstrip": false,
@@ -104,7 +104,7 @@
     },
     {
       "id": 11,
-      "content": "[control_9]",
+      "content": "<|im_end|>",
       "single_word": false,
       "lstrip": false,
       "rstrip": false,
@@ -113,7 +113,7 @@
     },
     {
       "id": 12,
-      "content": "[control_10]",
+      "content": "<|tools_start|>",
       "single_word": false,
       "lstrip": false,
       "rstrip": false,
@@ -122,7 +122,7 @@
     },
     {
       "id": 13,
-      "content": "[control_11]",
+      "content": "<|tools_end|>",
       "single_word": false,
       "lstrip": false,
       "rstrip": false,
@@ -131,7 +131,7 @@
     },
     {
       "id": 14,
-      "content": "[control_12]",
+      "content": "<|tool_start|>",
       "single_word": false,
       "lstrip": false,
       "rstrip": false,
@@ -140,7 +140,7 @@
     },
     {
       "id": 15,
-      "content": "[control_13]",
+      "content": "<|tool_end|>",
       "single_word": false,
       "lstrip": false,
       "rstrip": false,
@@ -149,7 +149,7 @@
     },
     {
       "id": 16,
-      "content": "[control_14]",
+      "content": "<|thinking|>",
       "single_word": false,
       "lstrip": false,
       "rstrip": false,
@@ -158,7 +158,7 @@
     },
     {
       "id": 17,
-      "content": "[control_15]",
+      "content": "<|thinking_end|>",
       "single_word": false,
       "lstrip": false,
       "rstrip": false,
@@ -7048,14 +7048,14 @@
       "[/AVAILABLE_TOOLS]": 7,
       "[TOOL_RESULTS]": 8,
       "[/TOOL_RESULTS]": 9,
-      "[control_8]": 10,
-      "[control_9]": 11,
-      "[control_10]": 12,
-      "[control_11]": 13,
-      "[control_12]": 14,
-      "[control_13]": 15,
-      "[control_14]": 16,
-      "[control_15]": 17,
+      "<|im_start|>": 10,
+      "<|im_end|>": 11,
+      "<|tools_start|>": 12,
+      "<|tools_end|>": 13,
+      "<|tool_start|>": 14,
+      "<|tool_end|>": 15,
+      "<|thinking|>": 16,
+      "<|thinking_end|>": 17,
       "[control_16]": 18,
       "[control_17]": 19,
       "[control_18]": 20,

tokenizer.model

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:37f00374dea48658ee8f5d0f21895b9bc55cb0103939607c8185bfd1c6ca1f89
-size 587404
+oid sha256:93e01cb5f215abfe797344a3b7bf0e874b49708177a8b115bbdde722983e3948
+size 587414

tokenizer_config.json

@@ -84,7 +84,7 @@
       "special": true
     },
     "10": {
-      "content": "[control_8]",
+      "content": "<|im_start|>",
       "lstrip": false,
       "normalized": false,
       "rstrip": false,
@@ -92,7 +92,7 @@
       "special": true
     },
     "11": {
-      "content": "[control_9]",
+      "content": "<|im_end|>",
       "lstrip": false,
       "normalized": false,
       "rstrip": false,
@@ -100,7 +100,7 @@
       "special": true
     },
     "12": {
-      "content": "[control_10]",
+      "content": "<|tools_start|>",
       "lstrip": false,
       "normalized": false,
       "rstrip": false,
@@ -108,7 +108,7 @@
       "special": true
     },
     "13": {
-      "content": "[control_11]",
+      "content": "<|tools_end|>",
       "lstrip": false,
       "normalized": false,
       "rstrip": false,
@@ -116,7 +116,7 @@
       "special": true
     },
     "14": {
-      "content": "[control_12]",
+      "content": "<|tool_start|>",
       "lstrip": false,
       "normalized": false,
       "rstrip": false,
@@ -124,7 +124,7 @@
       "special": true
     },
     "15": {
-      "content": "[control_13]",
+      "content": "<|tool_end|>",
       "lstrip": false,
       "normalized": false,
       "rstrip": false,
@@ -132,7 +132,7 @@
       "special": true
     },
     "16": {
-      "content": "[control_14]",
+      "content": "<|thinking|>",
       "lstrip": false,
       "normalized": false,
       "rstrip": false,
@@ -140,7 +140,7 @@
       "special": true
     },
     "17": {
-      "content": "[control_15]",
+      "content": "<|thinking_end|>",
       "lstrip": false,
       "normalized": false,
       "rstrip": false,
@@ -6174,7 +6174,7 @@
   },
   "bos_token": "<s>",
   "clean_up_tokenization_spaces": false,
-  "eos_token": "</s>",
+  "eos_token": "<|im_end|>",
   "legacy": false,
   "model_max_length": 1000000000000000019884624838656,
   "pad_token": "[control_768]",