Upload Kaggle-trained coding LoRA adapter

.gitattributes

@@ -33,3 +33,4 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
 *.zip filter=lfs diff=lfs merge=lfs -text
 *.zst filter=lfs diff=lfs merge=lfs -text
 *tfevents* filter=lfs diff=lfs merge=lfs -text
+tokenizer.json filter=lfs diff=lfs merge=lfs -text

README.md

@@ -0,0 +1,43 @@
+---
+base_model: google/gemma-4-E4B-it
+library_name: peft
+license: apache-2.0
+tags:
+- gemma4
+- coding
+- qlora
+- kaggle-proof
+---
+
+# Gemma 4 E4B IT Coding LoRA
+
+QLoRA adapter for `google/gemma-4-E4B-it`, trained on filtered benign coding instructions.
+
+## Training
+
+- Runtime: Kaggle 2x Tesla T4
+- Dataset: `ise-uiuc/Magicoder-Evol-Instruct-110K`, filtered to remove unsafe coding domains
+- Safe rows used: 1024
+- Steps: 200
+- LoRA: r=16, alpha=32, target_modules=`all-linear`
+- Trainable parameters: 50,499,584
+- Final train loss: 1.1427
+
+## Proof
+
+- HumanEval subset: first 8 tasks
+- Executable pass count before: 5/8
+- Executable pass count after: 7/8
+- Heuristic score before: 0.7688
+- Heuristic score after: 0.7688
+
+Artifacts included:
+
+- `eval_before_after.csv`
+- `executable_eval.json`
+- `trainer_log_history.json`
+- `summary.json`
+- `proof_summary.json`
+- `nvidia_smi.txt`
+
+This adapter is for benign coding assistance only. It was not trained on malware, phishing, exploit, credential theft, evasion, or destructive automation examples.

adapter_config.json

@@ -0,0 +1,308 @@
+{
+  "alora_invocation_tokens": null,
+  "alpha_pattern": {},
+  "arrow_config": null,
+  "auto_mapping": null,
+  "base_model_name_or_path": "google/gemma-4-E4B-it",
+  "bias": "none",
+  "corda_config": null,
+  "ensure_weight_tying": false,
+  "eva_config": null,
+  "exclude_modules": null,
+  "fan_in_fan_out": false,
+  "inference_mode": true,
+  "init_lora_weights": true,
+  "layer_replication": null,
+  "layers_pattern": null,
+  "layers_to_transform": null,
+  "loftq_config": {},
+  "lora_alpha": 32,
+  "lora_bias": false,
+  "lora_dropout": 0.05,
+  "lora_ga_config": null,
+  "megatron_config": null,
+  "megatron_core": "megatron.core",
+  "modules_to_save": null,
+  "peft_type": "LORA",
+  "peft_version": "0.19.1",
+  "qalora_group_size": 16,
+  "r": 16,
+  "rank_pattern": {},
+  "revision": null,
+  "target_modules": [
+    "input_proj_linear",
+    "input_proj",
+    "26.mlp.down_proj",
+    "30.mlp.gate_proj",
+    "language_model.layers.11.self_attn.q_proj",
+    "language_model.layers.13.mlp.down_proj",
+    "language_model.layers.11.mlp.down_proj",
+    "language_model.layers.14.mlp.down_proj",
+    "21.self_attn.o_proj",
+    "language_model.layers.3.mlp.up_proj",
+    "33.mlp.up_proj",
+    "language_model.layers.1.self_attn.v_proj",
+    "language_model.layers.13.self_attn.v_proj",
+    "language_model.layers.8.self_attn.v_proj",
+    "23.self_attn.k_proj",
+    "41.mlp.up_proj",
+    "28.self_attn.o_proj",
+    "22.self_attn.q_proj",
+    "language_model.layers.8.mlp.gate_proj",
+    "19.self_attn.o_proj",
+    "language_model.layers.15.self_attn.v_proj",
+    "language_model.layers.9.self_attn.q_proj",
+    "language_model.layers.4.self_attn.q_proj",
+    "35.self_attn.o_proj",
+    "linear",
+    "22.self_attn.o_proj",
+    "24.mlp.down_proj",
+    "31.mlp.up_proj",
+    "language_model.layers.10.self_attn.v_proj",
+    "20.mlp.gate_proj",
+    "27.mlp.up_proj",
+    "41.mlp.down_proj",
+    "18.self_attn.o_proj",
+    "language_model.layers.7.self_attn.q_proj",
+    "23.mlp.gate_proj",
+    "18.self_attn.k_proj",
+    "language_model.layers.9.mlp.down_proj",
+    "language_model.layers.3.self_attn.o_proj",
+    "language_model.layers.4.self_attn.o_proj",
+    "27.mlp.down_proj",
+    "16.mlp.up_proj",
+    "19.self_attn.k_proj",
+    "language_model.layers.14.mlp.gate_proj",
+    "language_model.layers.8.self_attn.q_proj",
+    "29.self_attn.q_proj",
+    "17.self_attn.v_proj",
+    "language_model.layers.8.mlp.up_proj",
+    "32.mlp.up_proj",
+    "language_model.layers.2.self_attn.v_proj",
+    "20.self_attn.q_proj",
+    "17.self_attn.q_proj",
+    "29.mlp.down_proj",
+    "language_model.layers.7.mlp.gate_proj",
+    "28.mlp.gate_proj",
+    "18.mlp.up_proj",
+    "language_model.layers.8.mlp.down_proj",
+    "37.self_attn.o_proj",
+    "language_model.layers.6.self_attn.o_proj",
+    "language_model.layers.11.mlp.up_proj",
+    "21.mlp.down_proj",
+    "language_model.layers.0.mlp.gate_proj",
+    "25.self_attn.o_proj",
+    "language_model.layers.0.self_attn.q_proj",
+    "language_model.layers.0.self_attn.o_proj",
+    "33.self_attn.q_proj",
+    "per_layer_model_projection",
+    "32.self_attn.q_proj",
+    "language_model.layers.7.self_attn.v_proj",
+    "41.self_attn.o_proj",
+    "38.mlp.down_proj",
+    "25.mlp.up_proj",
+    "23.self_attn.v_proj",
+    "26.self_attn.q_proj",
+    "16.self_attn.v_proj",
+    "language_model.layers.11.self_attn.v_proj",
+    "language_model.layers.10.self_attn.q_proj",
+    "language_model.layers.9.self_attn.v_proj",
+    "35.mlp.down_proj",
+    "language_model.layers.7.mlp.down_proj",
+    "language_model.layers.3.self_attn.k_proj",
+    "38.mlp.gate_proj",
+    "language_model.layers.4.self_attn.k_proj",
+    "37.self_attn.q_proj",
+    "language_model.layers.1.mlp.gate_proj",
+    "language_model.layers.13.self_attn.q_proj",
+    "40.mlp.gate_proj",
+    "language_model.layers.1.self_attn.o_proj",
+    "38.self_attn.q_proj",
+    "19.mlp.gate_proj",
+    "36.mlp.gate_proj",
+    "language_model.layers.15.mlp.gate_proj",
+    "language_model.layers.10.mlp.gate_proj",
+    "language_model.layers.12.self_attn.q_proj",
+    "language_model.layers.12.mlp.down_proj",
+    "16.self_attn.q_proj",
+    "21.self_attn.q_proj",
+    "language_model.layers.13.mlp.up_proj",
+    "36.mlp.up_proj",
+    "language_model.layers.9.mlp.up_proj",
+    "16.mlp.down_proj",
+    "language_model.layers.15.self_attn.o_proj",
+    "41.mlp.gate_proj",
+    "26.mlp.up_proj",
+    "30.mlp.down_proj",
+    "39.mlp.up_proj",
+    "21.mlp.gate_proj",
+    "language_model.layers.3.self_attn.q_proj",
+    "language_model.layers.0.mlp.down_proj",
+    "language_model.layers.14.self_attn.o_proj",
+    "language_model.layers.11.self_attn.o_proj",
+    "17.mlp.up_proj",
+    "29.mlp.up_proj",
+    "23.self_attn.q_proj",
+    "21.mlp.up_proj",
+    "39.mlp.down_proj",
+    "40.self_attn.o_proj",
+    "35.mlp.up_proj",
+    "language_model.layers.5.mlp.gate_proj",
+    "output_proj",
+    "22.mlp.gate_proj",
+    "30.self_attn.o_proj",
+    "language_model.layers.1.self_attn.k_proj",
+    "language_model.layers.10.self_attn.k_proj",
+    "language_model.layers.10.mlp.down_proj",
+    "31.mlp.down_proj",
+    "26.mlp.gate_proj",
+    "language_model.layers.0.self_attn.k_proj",
+    "per_layer_input_gate",
+    "17.mlp.down_proj",
+    "31.mlp.gate_proj",
+    "language_model.layers.2.self_attn.q_proj",
+    "24.self_attn.q_proj",
+    "22.mlp.up_proj",
+    "language_model.layers.4.mlp.up_proj",
+    "25.mlp.gate_proj",
+    "language_model.layers.5.mlp.down_proj",
+    "21.self_attn.k_proj",
+    "16.self_attn.o_proj",
+    "language_model.layers.5.self_attn.o_proj",
+    "19.mlp.up_proj",
+    "21.self_attn.v_proj",
+    "language_model.layers.10.mlp.up_proj",
+    "18.self_attn.q_proj",
+    "language_model.layers.10.self_attn.o_proj",
+    "16.self_attn.k_proj",
+    "language_model.layers.6.mlp.up_proj",
+    "language_model.layers.7.self_attn.o_proj",
+    "language_model.layers.15.self_attn.k_proj",
+    "language_model.layers.2.mlp.gate_proj",
+    "17.mlp.gate_proj",
+    "28.self_attn.q_proj",
+    "30.mlp.up_proj",
+    "language_model.layers.4.self_attn.v_proj",
+    "language_model.layers.4.mlp.down_proj",
+    "25.self_attn.q_proj",
+    "30.self_attn.q_proj",
+    "language_model.layers.5.self_attn.q_proj",
+    "language_model.layers.14.self_attn.v_proj",
+    "language_model.layers.3.mlp.down_proj",
+    "24.mlp.gate_proj",
+    "38.mlp.up_proj",
+    "language_model.layers.13.mlp.gate_proj",
+    "22.mlp.down_proj",
+    "39.mlp.gate_proj",
+    "35.mlp.gate_proj",
+    "17.self_attn.k_proj",
+    "embedding_projection",
+    "language_model.layers.0.mlp.up_proj",
+    "20.self_attn.o_proj",
+    "language_model.layers.14.self_attn.k_proj",
+    "20.mlp.up_proj",
+    "22.self_attn.v_proj",
+    "18.mlp.down_proj",
+    "language_model.layers.1.mlp.up_proj",
+    "39.self_attn.o_proj",
+    "31.self_attn.q_proj",
+    "34.mlp.gate_proj",
+    "language_model.layers.4.mlp.gate_proj",
+    "40.mlp.down_proj",
+    "36.self_attn.o_proj",
+    "27.self_attn.q_proj",
+    "34.mlp.up_proj",
+    "language_model.layers.5.self_attn.v_proj",
+    "language_model.layers.6.mlp.down_proj",
+    "language_model.layers.12.mlp.gate_proj",
+    "28.mlp.down_proj",
+    "29.self_attn.o_proj",
+    "language_model.layers.1.self_attn.q_proj",
+    "36.mlp.down_proj",
+    "language_model.layers.15.mlp.down_proj",
+    "32.self_attn.o_proj",
+    "40.mlp.up_proj",
+    "33.mlp.down_proj",
+    "26.self_attn.o_proj",
+    "language_model.layers.3.self_attn.v_proj",
+    "language_model.layers.2.self_attn.o_proj",
+    "language_model.layers.2.mlp.up_proj",
+    "language_model.layers.12.mlp.up_proj",
+    "language_model.layers.2.self_attn.k_proj",
+    "language_model.layers.12.self_attn.o_proj",
+    "language_model.layers.6.self_attn.k_proj",
+    "18.mlp.gate_proj",
+    "34.self_attn.o_proj",
+    "23.mlp.up_proj",
+    "29.mlp.gate_proj",
+    "language_model.layers.0.self_attn.v_proj",
+    "language_model.layers.15.mlp.up_proj",
+    "19.self_attn.q_proj",
+    "32.mlp.gate_proj",
+    "per_layer_projection",
+    "34.self_attn.q_proj",
+    "language_model.layers.1.mlp.down_proj",
+    "language_model.layers.13.self_attn.o_proj",
+    "language_model.layers.6.mlp.gate_proj",
+    "language_model.layers.3.mlp.gate_proj",
+    "20.mlp.down_proj",
+    "language_model.layers.2.mlp.down_proj",
+    "37.mlp.down_proj",
+    "language_model.layers.12.self_attn.k_proj",
+    "28.mlp.up_proj",
+    "37.mlp.gate_proj",
+    "27.mlp.gate_proj",
+    "language_model.layers.9.self_attn.k_proj",
+    "24.mlp.up_proj",
+    "20.self_attn.v_proj",
+    "language_model.layers.13.self_attn.k_proj",
+    "19.mlp.down_proj",
+    "23.mlp.down_proj",
+    "31.self_attn.o_proj",
+    "language_model.layers.5.self_attn.k_proj",
+    "17.self_attn.o_proj",
+    "16.mlp.gate_proj",
+    "39.self_attn.q_proj",
+    "language_model.layers.5.mlp.up_proj",
+    "language_model.layers.12.self_attn.v_proj",
+    "language_model.layers.14.self_attn.q_proj",
+    "36.self_attn.q_proj",
+    "24.self_attn.o_proj",
+    "23.self_attn.o_proj",
+    "language_model.layers.9.mlp.gate_proj",
+    "relative_k_proj",
+    "40.self_attn.q_proj",
+    "18.self_attn.v_proj",
+    "language_model.layers.11.self_attn.k_proj",
+    "32.mlp.down_proj",
+    "33.self_attn.o_proj",
+    "38.self_attn.o_proj",
+    "34.mlp.down_proj",
+    "language_model.layers.9.self_attn.o_proj",
+    "35.self_attn.q_proj",
+    "19.self_attn.v_proj",
+    "language_model.layers.8.self_attn.k_proj",
+    "33.mlp.gate_proj",
+    "language_model.layers.7.self_attn.k_proj",
+    "22.self_attn.k_proj",
+    "41.self_attn.q_proj",
+    "37.mlp.up_proj",
+    "27.self_attn.o_proj",
+    "20.self_attn.k_proj",
+    "language_model.layers.11.mlp.gate_proj",
+    "language_model.layers.14.mlp.up_proj",
+    "language_model.layers.8.self_attn.o_proj",
+    "language_model.layers.15.self_attn.q_proj",
+    "language_model.layers.6.self_attn.q_proj",
+    "language_model.layers.6.self_attn.v_proj",
+    "25.mlp.down_proj",
+    "language_model.layers.7.mlp.up_proj"
+  ],
+  "target_parameters": null,
+  "task_type": "CAUSAL_LM",
+  "trainable_token_indices": null,
+  "use_bdlora": null,
+  "use_dora": false,
+  "use_qalora": false,
+  "use_rslora": false
+}

adapter_model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:7225065dc08e4127efff0a65cfbad75b9c45e317897e1725e1c85bf07a53bda9
+size 202180544

chat_template.jinja

@@ -0,0 +1,351 @@
+{%- macro format_parameters(properties, required, filter_keys=false) -%}
+    {%- set standard_keys = ['description', 'type', 'properties', 'required', 'nullable'] -%}
+    {%- set ns = namespace(found_first=false) -%}
+    {%- for key, value in properties | dictsort -%}
+        {%- set add_comma = false -%}
+        {%- if not filter_keys or key not in standard_keys -%}
+            {%- if ns.found_first %},{% endif -%}
+            {%- set ns.found_first = true -%}
+            {{ key }}:{
+            {%- if value['description'] -%}
+                description:<|"|>{{ value['description'] }}<|"|>
+                {%- set add_comma = true -%}
+            {%- endif -%}
+            {%- if value['type'] | upper == 'STRING' -%}
+                {%- if value['enum'] -%}
+                    {%- if add_comma %},{%- else -%} {%- set add_comma = true -%} {% endif -%}
+                    enum:{{ format_argument(value['enum']) }}
+                {%- endif -%}
+            {%- elif value['type'] | upper == 'ARRAY' -%}
+                {%- if value['items'] is mapping and value['items'] -%}
+                    {%- if add_comma %},{%- else -%} {%- set add_comma = true -%} {% endif -%}
+                    items:{
+                    {%- set ns_items = namespace(found_first=false) -%}
+                    {%- for item_key, item_value in value['items'] | dictsort -%}
+                        {%- if item_value is not none -%}
+                            {%- if ns_items.found_first %},{% endif -%}
+                            {%- set ns_items.found_first = true -%}
+                            {%- if item_key == 'properties' -%}
+                                properties:{
+                                {%- if item_value is mapping -%}
+                                    {{- format_parameters(item_value, value['items']['required'] | default([])) -}}
+                                {%- endif -%}
+                                }
+                            {%- elif item_key == 'required' -%}
+                                required:[
+                                {%- for req_item in item_value -%}
+                                    <|"|>{{- req_item -}}<|"|>
+                                    {%- if not loop.last %},{% endif -%}
+                                {%- endfor -%}
+                                ]
+                            {%- elif item_key == 'type' -%}
+                                {%- if item_value is string -%}
+                                    type:{{ format_argument(item_value | upper) }}
+                                {%- else -%}
+                                    type:{{ format_argument(item_value | map('upper') | list) }}
+                                {%- endif -%}
+                            {%- else -%}
+                                {{ item_key }}:{{ format_argument(item_value) }}
+                            {%- endif -%}
+                        {%- endif -%}
+                    {%- endfor -%}
+                    }
+                {%- endif -%}
+            {%- endif -%}
+            {%- if value['nullable'] %}
+                {%- if add_comma %},{%- else -%} {%- set add_comma = true -%} {% endif -%}
+                nullable:true
+            {%- endif -%}
+            {%- if value['type'] | upper == 'OBJECT' -%}
+                {%- if value['properties'] is defined and value['properties'] is mapping -%}
+                    {%- if add_comma %},{%- else -%} {%- set add_comma = true -%} {% endif -%}
+                    properties:{
+                    {{- format_parameters(value['properties'], value['required'] | default([])) -}}
+                    }
+                {%- elif value is mapping -%}
+                    {%- if add_comma %},{%- else -%} {%- set add_comma = true -%} {% endif -%}
+                    properties:{
+                    {{- format_parameters(value, value['required'] | default([]), filter_keys=true) -}}
+                    }
+                {%- endif -%}
+                {%- if value['required'] -%}
+                    {%- if add_comma %},{%- else -%} {%- set add_comma = true -%} {% endif -%}
+                    required:[
+                    {%- for item in value['required'] | default([]) -%}
+                        <|"|>{{- item -}}<|"|>
+                        {%- if not loop.last %},{% endif -%}
+                    {%- endfor -%}
+                    ]
+                {%- endif -%}
+            {%- endif -%}
+            {%- if add_comma %},{%- else -%} {%- set add_comma = true -%} {% endif -%}
+            type:<|"|>{{ value['type'] | upper }}<|"|>}
+        {%- endif -%}
+    {%- endfor -%}
+{%- endmacro -%}
+{%- macro format_function_declaration(tool_data) -%}
+    declaration:{{- tool_data['function']['name'] -}}{description:<|"|>{{- tool_data['function']['description'] -}}<|"|>
+    {%- set params = tool_data['function']['parameters'] -%}
+    {%- if params -%}
+        ,parameters:{
+        {%- if params['properties'] -%}
+            properties:{ {{- format_parameters(params['properties'], params['required']) -}} },
+        {%- endif -%}
+        {%- if params['required'] -%}
+            required:[
+            {%- for item in params['required'] -%}
+                <|"|>{{- item -}}<|"|>
+                {{- ',' if not loop.last -}}
+            {%- endfor -%}
+            ],
+        {%- endif -%}
+        {%- if params['type'] -%}
+            type:<|"|>{{- params['type'] | upper -}}<|"|>}
+        {%- endif -%}
+    {%- endif -%}
+    {%- if 'response' in tool_data['function'] -%}
+        {%- set response_declaration = tool_data['function']['response'] -%}
+        ,response:{
+        {%- if response_declaration['description'] -%}
+            description:<|"|>{{- response_declaration['description'] -}}<|"|>,
+        {%- endif -%}
+        {%- if response_declaration['type'] | upper == 'OBJECT' -%}
+            type:<|"|>{{- response_declaration['type'] | upper -}}<|"|>}
+        {%- endif -%}
+    {%- endif -%}
+    }
+{%- endmacro -%}
+{%- macro format_argument(argument, escape_keys=True) -%}
+    {%- if argument is string -%}
+        {{- '<|"|>' + argument + '<|"|>' -}}
+    {%- elif argument is boolean -%}
+        {{- 'true' if argument else 'false' -}}
+    {%- elif argument is mapping -%}
+        {{- '{' -}}
+        {%- set ns = namespace(found_first=false) -%}
+        {%- for key, value in argument | dictsort -%}
+            {%- if ns.found_first %},{% endif -%}
+            {%- set ns.found_first = true -%}
+            {%- if escape_keys -%}
+                {{- '<|"|>' + key + '<|"|>' -}}
+            {%- else -%}
+                {{- key -}}
+            {%- endif -%}
+            :{{- format_argument(value, escape_keys=escape_keys) -}}
+        {%- endfor -%}
+        {{- '}' -}}
+    {%- elif argument is sequence -%}
+        {{- '[' -}}
+        {%- for item in argument -%}
+            {{- format_argument(item, escape_keys=escape_keys) -}}
+            {%- if not loop.last %},{% endif -%}
+        {%- endfor -%}
+        {{- ']' -}}
+    {%- else -%}
+        {{- argument -}}
+    {%- endif -%}
+{%- endmacro -%}
+{%- macro strip_thinking(text) -%}
+    {%- set ns = namespace(result='') -%}
+    {%- for part in text.split('<channel|>') -%}
+        {%- if '<|channel>' in part -%}
+            {%- set ns.result = ns.result + part.split('<|channel>')[0] -%}
+        {%- else -%}
+            {%- set ns.result = ns.result + part -%}
+        {%- endif -%}
+    {%- endfor -%}
+    {{- ns.result | trim -}}
+{%- endmacro -%}
+
+{%- macro format_tool_response_block(tool_name, response) -%}
+    {{- '<|tool_response>' -}}
+    {%- if response is mapping -%}
+        {{- 'response:' + tool_name + '{' -}}
+        {%- for key, value in response | dictsort -%}
+            {{- key -}}:{{- format_argument(value, escape_keys=False) -}}
+            {%- if not loop.last %},{% endif -%}
+        {%- endfor -%}
+        {{- '}' -}}
+    {%- else -%}
+        {{- 'response:' + tool_name + '{value:' + format_argument(response, escape_keys=False) + '}' -}}
+    {%- endif -%}
+    {{- '<tool_response|>' -}}
+{%- endmacro -%}
+
+{%- set ns = namespace(prev_message_type=None) -%}
+{%- set loop_messages = messages -%}
+{{- bos_token -}}
+{#- Handle System/Tool Definitions Block -#}
+{%- if (enable_thinking is defined and enable_thinking) or tools or messages[0]['role'] in ['system', 'developer'] -%}
+    {{- '<|turn>system\n' -}}
+    {#- Inject Thinking token at the very top of the FIRST system turn -#}
+    {%- if enable_thinking is defined and enable_thinking -%}
+        {{- '<|think|>\n' -}}
+        {%- set ns.prev_message_type = 'think' -%}
+    {%- endif -%}
+    {%- if messages[0]['role'] in ['system', 'developer'] -%}
+        {%- if messages[0]['content'] is string -%}
+            {{- messages[0]['content'] | trim -}}
+        {%- elif messages[0]['content'] is sequence -%}
+            {%- for item in messages[0]['content'] -%}
+                {{- item['text'] | trim + ' '-}}
+            {%- endfor -%}
+        {%- endif -%}
+        {%- set loop_messages = messages[1:] -%}
+    {%- endif -%}
+    {%- if tools -%}
+        {%- for tool in tools %}
+            {{- '<|tool>' -}}
+            {{- format_function_declaration(tool) | trim -}}
+            {{- '<tool|>' -}}
+        {%- endfor %}
+        {%- set ns.prev_message_type = 'tool' -%}
+    {%- endif -%}
+    {{- '<turn|>\n' -}}
+{%- endif %}
+
+{#- Pre-scan: find last user message index for reasoning guard -#}
+{%- set ns_turn = namespace(last_user_idx=-1) -%}
+{%- for i in range(loop_messages | length) -%}
+    {%- if loop_messages[i]['role'] == 'user' -%}
+        {%- set ns_turn.last_user_idx = i -%}
+    {%- endif -%}
+{%- endfor -%}
+
+{#- Loop through messages -#}
+{%- for message in loop_messages -%}
+    {%- if message['role'] != 'tool' -%}
+    {%- set ns.prev_message_type = None -%}
+    {%- set role = 'model' if message['role'] == 'assistant' else message['role'] -%}
+    {#- Detect continuation: suppress duplicate <|turn>model when previous non-tool message was also assistant -#}
+    {%- set prev_nt = namespace(role=None, found=false) -%}
+    {%- if loop.index0 > 0 -%}
+        {%- for j in range(loop.index0 - 1, -1, -1) -%}
+            {%- if not prev_nt.found -%}
+                {%- if loop_messages[j]['role'] != 'tool' -%}
+                    {%- set prev_nt.role = loop_messages[j]['role'] -%}
+                    {%- set prev_nt.found = true -%}
+                {%- endif -%}
+            {%- endif -%}
+        {%- endfor -%}
+    {%- endif -%}
+    {%- set continue_same_model_turn = (role == 'model' and prev_nt.role == 'assistant') -%}
+    {%- if not continue_same_model_turn -%}
+        {{- '<|turn>' + role + '\n' }}
+    {%- endif -%}
+
+    {#- Render reasoning/reasoning_content as thinking channel -#}
+    {%- set thinking_text = message.get('reasoning') or message.get('reasoning_content') -%}
+    {%- if thinking_text and loop.index0 > ns_turn.last_user_idx and message.get('tool_calls') -%}
+        {{- '<|channel>thought\n' + thinking_text + '\n<channel|>' -}}
+    {%- endif -%}
+
+            {%- if message['tool_calls'] -%}
+                {%- for tool_call in message['tool_calls'] -%}
+                    {%- set function = tool_call['function'] -%}
+                    {{- '<|tool_call>call:' + function['name'] + '{' -}}
+                    {%- if function['arguments'] is mapping -%}
+                        {%- set ns_args = namespace(found_first=false) -%}
+                        {%- for key, value in function['arguments'] | dictsort -%}
+                            {%- if ns_args.found_first %},{% endif -%}
+                            {%- set ns_args.found_first = true -%}
+                            {{- key -}}:{{- format_argument(value, escape_keys=False) -}}
+                        {%- endfor -%}
+                    {%- elif function['arguments'] is string -%}
+                        {{- function['arguments'] -}}
+                    {%- endif -%}
+                    {{- '}<tool_call|>' -}}
+                {%- endfor -%}
+                {%- set ns.prev_message_type = 'tool_call' -%}
+            {%- endif -%}
+
+            {%- set ns_tr_out = namespace(flag=false) -%}
+            {%- if message.get('tool_responses') -%}
+                {#- Legacy: tool_responses embedded on the assistant message (Google/Gemma native) -#}
+                {%- for tool_response in message['tool_responses'] -%}
+                    {{- format_tool_response_block(tool_response['name'] | default('unknown'), tool_response['response']) -}}
+                    {%- set ns_tr_out.flag = true -%}
+                    {%- set ns.prev_message_type = 'tool_response' -%}
+                {%- endfor -%}
+            {%- elif message.get('tool_calls') -%}
+                {#- OpenAI Chat Completions: forward-scan consecutive role:tool messages -#}
+                {%- set ns_tool_scan = namespace(stopped=false) -%}
+                {%- for k in range(loop.index0 + 1, loop_messages | length) -%}
+                    {%- if ns_tool_scan.stopped -%}
+                    {%- elif loop_messages[k]['role'] != 'tool' -%}
+                        {%- set ns_tool_scan.stopped = true -%}
+                    {%- else -%}
+                        {%- set follow = loop_messages[k] -%}
+                        {#- Resolve tool_call_id to function name -#}
+                        {%- set ns_tname = namespace(name=follow.get('name') | default('unknown')) -%}
+                        {%- for tc in message['tool_calls'] -%}
+                            {%- if tc.get('id') == follow.get('tool_call_id') -%}
+                                {%- set ns_tname.name = tc['function']['name'] -%}
+                            {%- endif -%}
+                        {%- endfor -%}
+                        {#- Handle content as string or content-parts array -#}
+                        {%- set tool_body = follow.get('content') -%}
+                        {%- if tool_body is string -%}
+                            {{- format_tool_response_block(ns_tname.name, tool_body) -}}
+                        {%- elif tool_body is sequence and tool_body is not string -%}
+                            {%- set ns_txt = namespace(s='') -%}
+                            {%- for part in tool_body -%}
+                                {%- if part.get('type') == 'text' -%}
+                                    {%- set ns_txt.s = ns_txt.s + (part.get('text') | default('')) -%}
+                                {%- endif -%}
+                            {%- endfor -%}
+                            {{- format_tool_response_block(ns_tname.name, ns_txt.s) -}}
+                        {%- else -%}
+                            {{- format_tool_response_block(ns_tname.name, tool_body) -}}
+                        {%- endif -%}
+                        {%- set ns_tr_out.flag = true -%}
+                        {%- set ns.prev_message_type = 'tool_response' -%}
+                    {%- endif -%}
+                {%- endfor -%}
+            {%- endif -%}
+
+            {%- set captured_content -%}
+            {%- if message['content'] is string -%}
+                {%- if role == 'model' -%}
+                    {{- strip_thinking(message['content']) -}}
+                {%- else -%}
+                    {{- message['content'] | trim -}}
+                {%- endif -%}
+            {%- elif message['content'] is sequence -%}
+                {%- for item in message['content'] -%}
+                    {%- if item['type'] == 'text' -%}
+                        {%- if role == 'model' -%}
+                            {{- strip_thinking(item['text']) -}}
+                        {%- else -%}
+                            {{- item['text'] | trim -}}
+                        {%- endif -%}
+                    {%- elif item['type'] == 'image' -%}
+                        {{- '<|image|>' -}}
+                        {%- set ns.prev_message_type = 'image' -%}
+                    {%- elif item['type'] == 'audio' -%}
+                        {{- '<|audio|>' -}}
+                        {%- set ns.prev_message_type = 'audio' -%}
+                    {%- elif item['type'] == 'video' -%}
+                        {{- '<|video|>' -}}
+                        {%- set ns.prev_message_type = 'video' -%}
+                    {%- endif -%}
+                {%- endfor -%}
+            {%- endif -%}
+            {%- endset -%}
+
+            {{- captured_content -}}
+            {%- set has_content = captured_content | trim | length > 0 -%}
+
+        {%- if ns.prev_message_type == 'tool_call' and not ns_tr_out.flag -%}
+            {{- '<|tool_response>' -}}
+        {%- elif not (ns_tr_out.flag and not has_content) -%}
+            {{- '<turn|>\n' -}}
+        {%- endif -%}
+    {%- endif -%}
+{%- endfor -%}
+
+{%- if add_generation_prompt -%}
+    {%- if ns.prev_message_type != 'tool_response' and ns.prev_message_type != 'tool_call' -%}
+        {{- '<|turn>model\n' -}}
+    {%- endif -%}
+{%- endif -%}

eval_before_after.csv

@@ -0,0 +1,150 @@
+phase,id,latency_s,score,chars,preview
+before,HumanEval_0,35.316,0.8,536,"```python
+from typing import List
+
+
+def has_close_elements(numbers: List[float], threshold: float) -> bool:
+    """""" Check if in given list of numbers, are any two numbers closer to each other than
+    given threshold.
+    >>> has_close_elements([1.0, 2.0, 3.0], 0.5)
+    False
+    >>> has_close_eleme"
+before,HumanEval_1,33.612,0.8,755,"```python
+from typing import List
+
+
+def separate_paren_groups(paren_string: str) -> List[str]:
+    """""" Input to this function is a string containing multiple groups of nested parentheses. Your goal is to
+    separate those group into separate strings and return the list of those.
+    Separate groups"
+before,HumanEval_2,17.675,0.8,377,"```python
+def truncate_number(number: float) -> float:
+    """""" Given a positive floating point number, it can be decomposed into
+    and integer part (largest integer smaller than given number) and decimals
+    (leftover part always smaller than 1).
+
+    Return the decimal part of the number.
+    >>"
+before,HumanEval_3,30.553,0.8,610,"```python
+from typing import List
+
+
+def below_zero(operations: List[int]) -> bool:
+    """""" You're given a list of deposit and withdrawal operations on a bank account that starts with
+    zero balance. Your task is to detect if at any point the balance of account fallls below zero, and
+    at that po"
+before,HumanEval_4,33.594,0.8,648,"```python
+from typing import List
+
+
+def mean_absolute_deviation(numbers: List[float]) -> float:
+    """""" For a given list of input numbers, calculate Mean Absolute Deviation
+    around the mean of this dataset.
+    Mean Absolute Deviation is the average absolute difference between each
+    element an"
+before,HumanEval_5,29.098,0.8,500,"```python
+from typing import List
+
+
+def intersperse(numbers: List[int], delimeter: int) -> List[int]:
+    """""" Insert a number 'delimeter' between every two consecutive elements of input list `numbers'
+    >>> intersperse([], 4)
+    []
+    >>> intersperse([1, 2, 3], 4)
+    [1, 4, 2, 4, 3]
+    """"""
+   "
+before,HumanEval_6,33.787,0.55,702,"```python
+from typing import List
+
+
+def parse_nested_parens(paren_string: str) -> List[int]:
+    """""" Input to this function is a string represented multiple groups for nested parentheses separated by spaces.
+    For each of the group, output the deepest level of nesting of parentheses.
+    E.g. (()("
+before,HumanEval_7,21.317,0.8,396,"```python
+from typing import List
+
+
+def filter_by_substring(strings: List[str], substring: str) -> List[str]:
+    """""" Filter an input list of strings only for ones that contain given substring
+    >>> filter_by_substring([], 'a')
+    []
+    >>> filter_by_substring(['abc', 'bacd', 'cde', 'array'], 'a"
+after,HumanEval_0,53.709,0.8,568,"```python
+from typing import List
+
+
+def has_close_elements(numbers: List[float], threshold: float) -> bool:
+    """""" Check if in given list of numbers, are any two numbers closer to each other than
+    given threshold.
+    >>> has_close_elements([1.0, 2.0, 3.0], 0.5)
+    False
+    >>> has_close_eleme"
+after,HumanEval_1,53.466,0.8,763,"```python
+from typing import List
+
+
+def separate_paren_groups(paren_string: str) -> List[str]:
+    paren_string = paren_string.replace("" "", """")
+    groups = []
+    balance = 0
+    start = 0
+    for i, char in enumerate(paren_string):
+        if char == '(':
+            balance += 1
+        elif char"
+after,HumanEval_2,53.294,0.8,824,"```python
+def truncate_number(number: float) -> float:
+    """""" Given a positive floating point number, it can be decomposed into
+    and an integer part (largest integer smaller than given number) and decimals
+    (leftover part always smaller than 1).
+
+    Return the decimal part of the number.
+   "
+after,HumanEval_3,53.551,0.8,770,"```python
+from typing import List
+
+
+def below_zero(operations: List[int]) -> bool:
+    """""" You're given a list of deposit and withdrawal operations on a bank account that starts with
+    zero balance. Your task is to detect if at any point the balance of account fallls below zero, and
+    at that po"
+after,HumanEval_4,53.816,0.8,736,"```python
+from typing import List
+
+
+def mean_absolute_deviation(numbers: List[float]) -> float:
+    """""" For a given list of input numbers, calculate Mean Absolute Deviation
+    around the mean of this dataset.
+    Mean Absolute Deviation is the average absolute difference between each
+    element an"
+after,HumanEval_5,53.518,0.8,643,"```python
+from typing import List
+
+
+def intersperse(numbers: List[int], delimeter: int) -> List[int]:
+    """""" Insert a number 'delimeter' between every two consecutive elements of input list `numbers'
+    >>> intersperse([], 4)
+    []
+    >>> intersperse([1, 2, 3], 4)
+    [1, 4, 2, 4, 3]
+    """"""
+   "
+after,HumanEval_6,53.528,0.55,701,"```python
+from typing import List
+
+
+def parse_nested_parens(paren_string: str) -> List[int]:
+    """""" Input to this function is a string represented multiple groups for nested parentheses separated by spaces.
+    For each of the group, output the deepest level of nesting of parentheses.
+    E.g. (()("
+after,HumanEval_7,53.317,0.8,739,"```python
+from typing import List
+
+
+def filter_by_substring(strings: List[str], substring: str) -> List[str]:
+    """""" Filter an input list of strings only for ones that contain given substring
+    >>> filter_by_substring([], 'a')
+    []
+    >>> filter_by_substring(['abc', 'bacd', 'cde', 'array'], 'a"

executable_eval.json

@@ -0,0 +1,135 @@
+{
+  "before_pass": 5,
+  "after_pass": 7,
+  "total": 8,
+  "rows": [
+    {
+      "phase": "before",
+      "task_id": "HumanEval/0",
+      "entry_point": "has_close_elements",
+      "passed": false,
+      "error": "Traceback (most recent call last):\n  File \"C:\\Users\\USER\\AppData\\Local\\Temp\\tmp6oy5omdq.py\", line 8, in <module>\n    exec(code, ns)\n    ~~~~^^^^^^^^^^\n  File \"<string>\", line 17\n    if sorted_numbers[i+1] - sorted_numbers[i\n                                           ^\nSyntaxError: '[' was never closed\n",
+      "chars": 526
+    },
+    {
+      "phase": "after",
+      "task_id": "HumanEval/0",
+      "entry_point": "has_close_elements",
+      "passed": true,
+      "error": null,
+      "chars": 495
+    },
+    {
+      "phase": "before",
+      "task_id": "HumanEval/1",
+      "entry_point": "separate_paren_groups",
+      "passed": false,
+      "error": "Traceback (most recent call last):\n  File \"C:\\Users\\USER\\AppData\\Local\\Temp\\tmpbgx0dlv4.py\", line 10, in <module>\n    ns[\"check\"](ns[entry_point])\n    ~~~~~~~~~~~^^^^^^^^^^^^^^^^^\n  File \"<string>\", line 10, in check\nAssertionError\n",
+      "chars": 745
+    },
+    {
+      "phase": "after",
+      "task_id": "HumanEval/1",
+      "entry_point": "separate_paren_groups",
+      "passed": true,
+      "error": null,
+      "chars": 455
+    },
+    {
+      "phase": "before",
+      "task_id": "HumanEval/2",
+      "entry_point": "truncate_number",
+      "passed": true,
+      "error": null,
+      "chars": 360
+    },
+    {
+      "phase": "after",
+      "task_id": "HumanEval/2",
+      "entry_point": "truncate_number",
+      "passed": true,
+      "error": null,
+      "chars": 363
+    },
+    {
+      "phase": "before",
+      "task_id": "HumanEval/3",
+      "entry_point": "below_zero",
+      "passed": true,
+      "error": null,
+      "chars": 590
+    },
+    {
+      "phase": "after",
+      "task_id": "HumanEval/3",
+      "entry_point": "below_zero",
+      "passed": true,
+      "error": null,
+      "chars": 590
+    },
+    {
+      "phase": "before",
+      "task_id": "HumanEval/4",
+      "entry_point": "mean_absolute_deviation",
+      "passed": true,
+      "error": null,
+      "chars": 632
+    },
+    {
+      "phase": "after",
+      "task_id": "HumanEval/4",
+      "entry_point": "mean_absolute_deviation",
+      "passed": true,
+      "error": null,
+      "chars": 530
+    },
+    {
+      "phase": "before",
+      "task_id": "HumanEval/5",
+      "entry_point": "intersperse",
+      "passed": true,
+      "error": null,
+      "chars": 486
+    },
+    {
+      "phase": "after",
+      "task_id": "HumanEval/5",
+      "entry_point": "intersperse",
+      "passed": true,
+      "error": null,
+      "chars": 455
+    },
+    {
+      "phase": "before",
+      "task_id": "HumanEval/6",
+      "entry_point": "parse_nested_parens",
+      "passed": false,
+      "error": "Traceback (most recent call last):\n  File \"C:\\Users\\USER\\AppData\\Local\\Temp\\tmpx0bb66c4.py\", line 8, in <module>\n    exec(code, ns)\n    ~~~~^^^^^^^^^^\n  File \"<string>\", line 21\n    max_depth = max(max_depth\n                   ^\nSyntaxError: '(' was never closed\n",
+      "chars": 692
+    },
+    {
+      "phase": "after",
+      "task_id": "HumanEval/6",
+      "entry_point": "parse_nested_parens",
+      "passed": false,
+      "error": "Traceback (most recent call last):\n  File \"C:\\Users\\USER\\AppData\\Local\\Temp\\tmpn7um62g2.py\", line 8, in <module>\n    exec(code, ns)\n    ~~~~^^^^^^^^^^\n  File \"<string>\", line 20\n    max_depth = max(max_depth\n                   ^\nSyntaxError: '(' was never closed\n",
+      "chars": 691
+    },
+    {
+      "phase": "before",
+      "task_id": "HumanEval/7",
+      "entry_point": "filter_by_substring",
+      "passed": true,
+      "error": null,
+      "chars": 379
+    },
+    {
+      "phase": "after",
+      "task_id": "HumanEval/7",
+      "entry_point": "filter_by_substring",
+      "passed": true,
+      "error": null,
+      "chars": 379
+    }
+  ]
+}

nvidia_smi.txt

@@ -0,0 +1,24 @@
+Tue May 12 19:04:42 2026       
++-----------------------------------------------------------------------------------------+
+| NVIDIA-SMI 580.105.08             Driver Version: 580.105.08     CUDA Version: 13.0     |
++-----------------------------------------+------------------------+----------------------+
+| GPU  Name                 Persistence-M | Bus-Id          Disp.A | Volatile Uncorr. ECC |
+| Fan  Temp   Perf          Pwr:Usage/Cap |           Memory-Usage | GPU-Util  Compute M. |
+|                                         |                        |               MIG M. |
+|=========================================+========================+======================|
+|   0  Tesla T4                       Off |   00000000:00:04.0 Off |                    0 |
+| N/A   39C    P8              9W /   70W |       3MiB /  15360MiB |      0%      Default |
+|                                         |                        |                  N/A |
++-----------------------------------------+------------------------+----------------------+
+|   1  Tesla T4                       Off |   00000000:00:05.0 Off |                    0 |
+| N/A   38C    P8             10W /   70W |       3MiB /  15360MiB |      0%      Default |
+|                                         |                        |                  N/A |
++-----------------------------------------+------------------------+----------------------+
+
++-----------------------------------------------------------------------------------------+
+| Processes:                                                                              |
+|  GPU   GI   CI              PID   Type   Process name                        GPU Memory |
+|        ID   ID                                                               Usage      |
+|=========================================================================================|
+|  No running processes found                                                             |
++-----------------------------------------------------------------------------------------+

processor_config.json

@@ -0,0 +1,75 @@
+{
+  "audio_ms_per_token": 40,
+  "audio_seq_length": 750,
+  "feature_extractor": {
+    "dither": 0.0,
+    "feature_extractor_type": "Gemma4AudioFeatureExtractor",
+    "feature_size": 128,
+    "fft_length": 512,
+    "fft_overdrive": false,
+    "frame_length": 320,
+    "hop_length": 160,
+    "input_scale_factor": 1.0,
+    "max_frequency": 8000.0,
+    "mel_floor": 0.001,
+    "min_frequency": 0.0,
+    "padding_side": "right",
+    "padding_value": 0.0,
+    "per_bin_mean": null,
+    "per_bin_stddev": null,
+    "preemphasis": 0.0,
+    "preemphasis_htk_flavor": true,
+    "return_attention_mask": true,
+    "sampling_rate": 16000
+  },
+  "image_processor": {
+    "do_convert_rgb": true,
+    "do_normalize": false,
+    "do_rescale": true,
+    "do_resize": true,
+    "image_mean": [
+      0.0,
+      0.0,
+      0.0
+    ],
+    "image_processor_type": "Gemma4ImageProcessor",
+    "image_seq_length": 280,
+    "image_std": [
+      1.0,
+      1.0,
+      1.0
+    ],
+    "max_soft_tokens": 280,
+    "patch_size": 16,
+    "pooling_kernel_size": 3,
+    "resample": 3,
+    "rescale_factor": 0.00392156862745098
+  },
+  "image_seq_length": 280,
+  "processor_class": "Gemma4Processor",
+  "video_processor": {
+    "do_convert_rgb": true,
+    "do_normalize": true,
+    "do_rescale": true,
+    "do_resize": true,
+    "do_sample_frames": true,
+    "image_mean": [
+      0.0,
+      0.0,
+      0.0
+    ],
+    "image_std": [
+      1.0,
+      1.0,
+      1.0
+    ],
+    "max_soft_tokens": 70,
+    "num_frames": 32,
+    "patch_size": 16,
+    "pooling_kernel_size": 3,
+    "resample": 3,
+    "rescale_factor": 0.00392156862745098,
+    "return_metadata": false,
+    "video_processor_type": "Gemma4VideoProcessor"
+  }
+}

proof_summary.json

@@ -0,0 +1,175 @@
+{
+  "run_id": "20260512_190433",
+  "model_id": "google/gemma-4-E4B-it",
+  "dataset_id": "ise-uiuc/Magicoder-Evol-Instruct-110K",
+  "adapter_repo": "josephmayo/gemma-4-E4B-it-coding-lora",
+  "merged_repo": "josephmayo/gemma-4-E4B-it-coding-merged",
+  "stage": "after_eval",
+  "errors": [],
+  "cuda_available": true,
+  "cuda_device_count": 2,
+  "devices": [
+    "Tesla T4",
+    "Tesla T4"
+  ],
+  "torch_version_initial": "2.10.0+cu128",
+  "hf_token_present": false,
+  "max_train_samples": 1024,
+  "max_steps": 200,
+  "max_seq_length": 512,
+  "eval_count": 8,
+  "lora_r": 16,
+  "lora_alpha": 32,
+  "lr": 0.0001,
+  "grad_accum": 8,
+  "push_to_hf": true,
+  "merge_and_push": false,
+  "load_in_4bit": true,
+  "memory_after_load": [
+    0,
+    9302143488
+  ],
+  "eval_source": "openai/openai_humaneval:8",
+  "baseline_avg_score": 0.76875,
+  "safe_train_rows": 1024,
+  "trainable_parameters": {
+    "trainable": 50499584,
+    "total": 7991600416
+  },
+  "log_history_tail": [
+    {
+      "loss": 1.0154043197631837,
+      "grad_norm": 0.37521687150001526,
+      "learning_rate": 4.51495073572676e-05,
+      "epoch": 1.71875
+    },
+    {
+      "loss": 0.9917967796325684,
+      "grad_norm": 0.42887604236602783,
+      "learning_rate": 4.114045042103887e-05,
+      "epoch": 1.796875
+    },
+    {
+      "loss": 1.1146905899047852,
+      "grad_norm": 0.4208148717880249,
+      "learning_rate": 3.718944461187138e-05,
+      "epoch": 1.875
+    },
+    {
+      "loss": 0.9283761978149414,
+      "grad_norm": 0.3849687874317169,
+      "learning_rate": 3.332237841745898e-05,
+      "epoch": 1.953125
+    },
+    {
+      "loss": 1.113053035736084,
+      "grad_norm": 0.4142734110355377,
+      "learning_rate": 2.9564590321322207e-05,
+      "epoch": 2.03125
+    },
+    {
+      "loss": 0.9842248916625976,
+      "grad_norm": 0.44529953598976135,
+      "learning_rate": 2.5940702775459747e-05,
+      "epoch": 2.109375
+    },
+    {
+      "loss": 0.9449721336364746,
+      "grad_norm": 0.3756776750087738,
+      "learning_rate": 2.2474460864709824e-05,
+      "epoch": 2.1875
+    },
+    {
+      "loss": 1.0590093612670899,
+      "grad_norm": 0.4192875325679779,
+      "learning_rate": 1.9188576719953633e-05,
+      "epoch": 2.265625
+    },
+    {
+      "loss": 0.9768091201782226,
+      "grad_norm": 0.5095818638801575,
+      "learning_rate": 1.6104580699624837e-05,
+      "epoch": 2.34375
+    },
+    {
+      "loss": 1.038302516937256,
+      "grad_norm": 0.41709497570991516,
+      "learning_rate": 1.3242680314639993e-05,
+      "epoch": 2.421875
+    },
+    {
+      "loss": 0.9975608825683594,
+      "grad_norm": 0.5563586354255676,
+      "learning_rate": 1.0621627821127289e-05,
+      "epoch": 2.5
+    },
+    {
+      "loss": 0.9714397430419922,
+      "grad_norm": 0.8915637135505676,
+      "learning_rate": 8.25859734853645e-06,
+      "epoch": 2.578125
+    },
+    {
+      "loss": 0.9948483467102051,
+      "grad_norm": 0.4391196370124817,
+      "learning_rate": 6.16907236823262e-06,
+      "epoch": 2.65625
+    },
+    {
+      "loss": 0.9389057159423828,
+      "grad_norm": 0.4650712311267853,
+      "learning_rate": 4.366744239922998e-06,
+      "epoch": 2.734375
+    },
+    {
+      "loss": 1.06390380859375,
+      "grad_norm": 0.4836062788963318,
+      "learning_rate": 2.8634225006782865e-06,
+      "epoch": 2.8125
+    },
+    {
+      "loss": 1.008359718322754,
+      "grad_norm": 0.45215511322021484,
+      "learning_rate": 1.6689574843694433e-06,
+      "epoch": 2.890625
+    },
+    {
+      "loss": 1.0110493659973145,
+      "grad_norm": 0.5408219695091248,
+      "learning_rate": 7.911757785462881e-07,
+      "epoch": 2.96875
+    },
+    {
+      "loss": 0.911649227142334,
+      "grad_norm": 0.4599083364009857,
+      "learning_rate": 2.3582894166930268e-07,
+      "epoch": 3.046875
+    },
+    {
+      "loss": 0.9673548698425293,
+      "grad_norm": 0.43304941058158875,
+      "learning_rate": 6.5558167183898955e-09,
+      "epoch": 3.125
+    },
+    {
+      "train_runtime": 4256.6409,
+      "train_samples_per_second": 0.752,
+      "train_steps_per_second": 0.047,
+      "total_flos": 4.259313762009523e+16,
+      "train_loss": 1.142699921131134,
+      "epoch": 3.125
+    }
+  ],
+  "train_metrics": {
+    "train_runtime": 4256.6409,
+    "train_samples_per_second": 0.752,
+    "train_steps_per_second": 0.047,
+    "total_flos": 4.259313762009523e+16,
+    "train_loss": 1.142699921131134,
+    "epoch": 3.125
+  },
+  "after_avg_score": 0.76875,
+  "score_delta": 0.0,
+  "adapter_dir": "/kaggle/working/gemma4_e4b_coding_lora",
+  "release_gate_pass": true
+}

summary.json

@@ -0,0 +1,175 @@
+{
+  "run_id": "20260512_190433",
+  "model_id": "google/gemma-4-E4B-it",
+  "dataset_id": "ise-uiuc/Magicoder-Evol-Instruct-110K",
+  "adapter_repo": "josephmayo/gemma-4-E4B-it-coding-lora",
+  "merged_repo": "josephmayo/gemma-4-E4B-it-coding-merged",
+  "stage": "after_eval",
+  "errors": [],
+  "cuda_available": true,
+  "cuda_device_count": 2,
+  "devices": [
+    "Tesla T4",
+    "Tesla T4"
+  ],
+  "torch_version_initial": "2.10.0+cu128",
+  "hf_token_present": false,
+  "max_train_samples": 1024,
+  "max_steps": 200,
+  "max_seq_length": 512,
+  "eval_count": 8,
+  "lora_r": 16,
+  "lora_alpha": 32,
+  "lr": 0.0001,
+  "grad_accum": 8,
+  "push_to_hf": true,
+  "merge_and_push": false,
+  "load_in_4bit": true,
+  "memory_after_load": [
+    0,
+    9302143488
+  ],
+  "eval_source": "openai/openai_humaneval:8",
+  "baseline_avg_score": 0.76875,
+  "safe_train_rows": 1024,
+  "trainable_parameters": {
+    "trainable": 50499584,
+    "total": 7991600416
+  },
+  "log_history_tail": [
+    {
+      "loss": 1.0154043197631837,
+      "grad_norm": 0.37521687150001526,
+      "learning_rate": 4.51495073572676e-05,
+      "epoch": 1.71875
+    },
+    {
+      "loss": 0.9917967796325684,
+      "grad_norm": 0.42887604236602783,
+      "learning_rate": 4.114045042103887e-05,
+      "epoch": 1.796875
+    },
+    {
+      "loss": 1.1146905899047852,
+      "grad_norm": 0.4208148717880249,
+      "learning_rate": 3.718944461187138e-05,
+      "epoch": 1.875
+    },
+    {
+      "loss": 0.9283761978149414,
+      "grad_norm": 0.3849687874317169,
+      "learning_rate": 3.332237841745898e-05,
+      "epoch": 1.953125
+    },
+    {
+      "loss": 1.113053035736084,
+      "grad_norm": 0.4142734110355377,
+      "learning_rate": 2.9564590321322207e-05,
+      "epoch": 2.03125
+    },
+    {
+      "loss": 0.9842248916625976,
+      "grad_norm": 0.44529953598976135,
+      "learning_rate": 2.5940702775459747e-05,
+      "epoch": 2.109375
+    },
+    {
+      "loss": 0.9449721336364746,
+      "grad_norm": 0.3756776750087738,
+      "learning_rate": 2.2474460864709824e-05,
+      "epoch": 2.1875
+    },
+    {
+      "loss": 1.0590093612670899,
+      "grad_norm": 0.4192875325679779,
+      "learning_rate": 1.9188576719953633e-05,
+      "epoch": 2.265625
+    },
+    {
+      "loss": 0.9768091201782226,
+      "grad_norm": 0.5095818638801575,
+      "learning_rate": 1.6104580699624837e-05,
+      "epoch": 2.34375
+    },
+    {
+      "loss": 1.038302516937256,
+      "grad_norm": 0.41709497570991516,
+      "learning_rate": 1.3242680314639993e-05,
+      "epoch": 2.421875
+    },
+    {
+      "loss": 0.9975608825683594,
+      "grad_norm": 0.5563586354255676,
+      "learning_rate": 1.0621627821127289e-05,
+      "epoch": 2.5
+    },
+    {
+      "loss": 0.9714397430419922,
+      "grad_norm": 0.8915637135505676,
+      "learning_rate": 8.25859734853645e-06,
+      "epoch": 2.578125
+    },
+    {
+      "loss": 0.9948483467102051,
+      "grad_norm": 0.4391196370124817,
+      "learning_rate": 6.16907236823262e-06,
+      "epoch": 2.65625
+    },
+    {
+      "loss": 0.9389057159423828,
+      "grad_norm": 0.4650712311267853,
+      "learning_rate": 4.366744239922998e-06,
+      "epoch": 2.734375
+    },
+    {
+      "loss": 1.06390380859375,
+      "grad_norm": 0.4836062788963318,
+      "learning_rate": 2.8634225006782865e-06,
+      "epoch": 2.8125
+    },
+    {
+      "loss": 1.008359718322754,
+      "grad_norm": 0.45215511322021484,
+      "learning_rate": 1.6689574843694433e-06,
+      "epoch": 2.890625
+    },
+    {
+      "loss": 1.0110493659973145,
+      "grad_norm": 0.5408219695091248,
+      "learning_rate": 7.911757785462881e-07,
+      "epoch": 2.96875
+    },
+    {
+      "loss": 0.911649227142334,
+      "grad_norm": 0.4599083364009857,
+      "learning_rate": 2.3582894166930268e-07,
+      "epoch": 3.046875
+    },
+    {
+      "loss": 0.9673548698425293,
+      "grad_norm": 0.43304941058158875,
+      "learning_rate": 6.5558167183898955e-09,
+      "epoch": 3.125
+    },
+    {
+      "train_runtime": 4256.6409,
+      "train_samples_per_second": 0.752,
+      "train_steps_per_second": 0.047,
+      "total_flos": 4.259313762009523e+16,
+      "train_loss": 1.142699921131134,
+      "epoch": 3.125
+    }
+  ],
+  "train_metrics": {
+    "train_runtime": 4256.6409,
+    "train_samples_per_second": 0.752,
+    "train_steps_per_second": 0.047,
+    "total_flos": 4.259313762009523e+16,
+    "train_loss": 1.142699921131134,
+    "epoch": 3.125
+  },
+  "after_avg_score": 0.76875,
+  "score_delta": 0.0,
+  "adapter_dir": "/kaggle/working/gemma4_e4b_coding_lora",
+  "release_gate_pass": true
+}

tokenizer.json

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:88e7140798a237085e49912b68c73b1928d746c7d263133d61f7c3f39dca8431
+size 32169724

tokenizer_config.json

@@ -0,0 +1,96 @@
+{
+  "audio_token": "<|audio|>",
+  "backend": "tokenizers",
+  "boa_token": "<|audio>",
+  "boi_token": "<|image>",
+  "bos_token": "<bos>",
+  "eoa_token": "<audio|>",
+  "eoc_token": "<channel|>",
+  "eoi_token": "<image|>",
+  "eos_token": "<eos>",
+  "eot_token": "<turn|>",
+  "escape_token": "<|\"|>",
+  "etc_token": "<tool_call|>",
+  "etd_token": "<tool|>",
+  "etr_token": "<tool_response|>",
+  "extra_special_tokens": [
+    "<|video|>"
+  ],
+  "image_token": "<|image|>",
+  "is_local": false,
+  "local_files_only": false,
+  "mask_token": "<mask>",
+  "model_max_length": 1000000000000000019884624838656,
+  "model_specific_special_tokens": {
+    "audio_token": "<|audio|>",
+    "boa_token": "<|audio>",
+    "boi_token": "<|image>",
+    "eoa_token": "<audio|>",
+    "eoc_token": "<channel|>",
+    "eoi_token": "<image|>",
+    "eot_token": "<turn|>",
+    "escape_token": "<|\"|>",
+    "etc_token": "<tool_call|>",
+    "etd_token": "<tool|>",
+    "etr_token": "<tool_response|>",
+    "image_token": "<|image|>",
+    "soc_token": "<|channel>",
+    "sot_token": "<|turn>",
+    "stc_token": "<|tool_call>",
+    "std_token": "<|tool>",
+    "str_token": "<|tool_response>",
+    "think_token": "<|think|>"
+  },
+  "pad_token": "<pad>",
+  "padding_side": "left",
+  "processor_class": "Gemma4Processor",
+  "response_schema": {
+    "properties": {
+      "content": {
+        "type": "string"
+      },
+      "role": {
+        "const": "assistant"
+      },
+      "thinking": {
+        "type": "string"
+      },
+      "tool_calls": {
+        "items": {
+          "properties": {
+            "function": {
+              "properties": {
+                "arguments": {
+                  "additionalProperties": {},
+                  "type": "object",
+                  "x-parser": "gemma4-tool-call"
+                },
+                "name": {
+                  "type": "string"
+                }
+              },
+              "type": "object",
+              "x-regex": "call\\:(?P<name>\\w+)(?P<arguments>\\{.*\\})"
+            },
+            "type": {
+              "const": "function"
+            }
+          },
+          "type": "object"
+        },
+        "type": "array",
+        "x-regex-iterator": "<\\|tool_call>(.*?)<tool_call\\|>"
+      }
+    },
+    "type": "object",
+    "x-regex": "(\\<\\|channel\\>thought\\n(?P<thinking>.*?)\\<channel\\|\\>)?(?P<tool_calls>\\<\\|tool_call\\>.*\\<tool_call\\|\\>)?(?P<content>(?:(?!\\<turn\\|\\>)(?!\\<\\|tool_response\\>).)+)?(?:\\<turn\\|\\>|\\<\\|tool_response\\>)?"
+  },
+  "soc_token": "<|channel>",
+  "sot_token": "<|turn>",
+  "stc_token": "<|tool_call>",
+  "std_token": "<|tool>",
+  "str_token": "<|tool_response>",
+  "think_token": "<|think|>",
+  "tokenizer_class": "GemmaTokenizer",
+  "unk_token": "<unk>"
+}

trainer_log_history.json

@@ -0,0 +1,291 @@
+[
+  {
+    "loss": 2.7602264404296877,
+    "grad_norm": 3.374150276184082,
+    "learning_rate": 6.666666666666667e-05,
+    "epoch": 0.078125,
+    "step": 5
+  },
+  {
+    "loss": 1.8698038101196288,
+    "grad_norm": 1.457783579826355,
+    "learning_rate": 9.994100796397954e-05,
+    "epoch": 0.15625,
+    "step": 10
+  },
+  {
+    "loss": 1.319937038421631,
+    "grad_norm": 0.707484245300293,
+    "learning_rate": 9.958100506132127e-05,
+    "epoch": 0.234375,
+    "step": 15
+  },
+  {
+    "loss": 1.3707428932189942,
+    "grad_norm": 0.43754199147224426,
+    "learning_rate": 9.889612861977853e-05,
+    "epoch": 0.3125,
+    "step": 20
+  },
+  {
+    "loss": 1.3082144737243653,
+    "grad_norm": 0.35036444664001465,
+    "learning_rate": 9.789086620939936e-05,
+    "epoch": 0.390625,
+    "step": 25
+  },
+  {
+    "loss": 1.283499526977539,
+    "grad_norm": 0.26873722672462463,
+    "learning_rate": 9.657180469054213e-05,
+    "epoch": 0.46875,
+    "step": 30
+  },
+  {
+    "loss": 1.1428126335144042,
+    "grad_norm": 0.31287550926208496,
+    "learning_rate": 9.494758705426978e-05,
+    "epoch": 0.546875,
+    "step": 35
+  },
+  {
+    "loss": 1.1817453384399415,
+    "grad_norm": 0.2988075911998749,
+    "learning_rate": 9.302885579019627e-05,
+    "epoch": 0.625,
+    "step": 40
+  },
+  {
+    "loss": 1.1872751235961914,
+    "grad_norm": 0.2937701642513275,
+    "learning_rate": 9.082818315286055e-05,
+    "epoch": 0.703125,
+    "step": 45
+  },
+  {
+    "loss": 1.146270751953125,
+    "grad_norm": 0.3410361111164093,
+    "learning_rate": 8.835998878354931e-05,
+    "epoch": 0.78125,
+    "step": 50
+  },
+  {
+    "loss": 1.167984962463379,
+    "grad_norm": 0.3269799053668976,
+    "learning_rate": 8.564044522734147e-05,
+    "epoch": 0.859375,
+    "step": 55
+  },
+  {
+    "loss": 1.0844226837158204,
+    "grad_norm": 0.44009652733802795,
+    "learning_rate": 8.268737196446264e-05,
+    "epoch": 0.9375,
+    "step": 60
+  },
+  {
+    "loss": 1.1493185043334961,
+    "grad_norm": 0.383281409740448,
+    "learning_rate": 7.952011865029614e-05,
+    "epoch": 1.015625,
+    "step": 65
+  },
+  {
+    "loss": 1.1071245193481445,
+    "grad_norm": 0.40226107835769653,
+    "learning_rate": 7.61594383291065e-05,
+    "epoch": 1.09375,
+    "step": 70
+  },
+  {
+    "loss": 1.062960433959961,
+    "grad_norm": 0.3419550359249115,
+    "learning_rate": 7.262735145222696e-05,
+    "epoch": 1.171875,
+    "step": 75
+  },
+  {
+    "loss": 1.1302300453186036,
+    "grad_norm": 0.39782464504241943,
+    "learning_rate": 6.894700159171534e-05,
+    "epoch": 1.25,
+    "step": 80
+  },
+  {
+    "loss": 1.0111728668212892,
+    "grad_norm": 0.3370531499385834,
+    "learning_rate": 6.514250379489753e-05,
+    "epoch": 1.328125,
+    "step": 85
+  },
+  {
+    "loss": 1.0343121528625487,
+    "grad_norm": 0.35490792989730835,
+    "learning_rate": 6.123878657343648e-05,
+    "epoch": 1.40625,
+    "step": 90
+  },
+  {
+    "loss": 1.161344337463379,
+    "grad_norm": 0.4397459328174591,
+    "learning_rate": 5.726142856227452e-05,
+    "epoch": 1.484375,
+    "step": 95
+  },
+  {
+    "loss": 1.1250411987304687,
+    "grad_norm": 0.375924289226532,
+    "learning_rate": 5.3236490918721794e-05,
+    "epoch": 1.5625,
+    "step": 100
+  },
+  {
+    "loss": 1.0718464851379395,
+    "grad_norm": 0.41714778542518616,
+    "learning_rate": 4.919034655987493e-05,
+    "epoch": 1.640625,
+    "step": 105
+  },
+  {
+    "loss": 1.0154043197631837,
+    "grad_norm": 0.37521687150001526,
+    "learning_rate": 4.51495073572676e-05,
+    "epoch": 1.71875,
+    "step": 110
+  },
+  {
+    "loss": 0.9917967796325684,
+    "grad_norm": 0.42887604236602783,
+    "learning_rate": 4.114045042103887e-05,
+    "epoch": 1.796875,
+    "step": 115
+  },
+  {
+    "loss": 1.1146905899047852,
+    "grad_norm": 0.4208148717880249,
+    "learning_rate": 3.718944461187138e-05,
+    "epoch": 1.875,
+    "step": 120
+  },
+  {
+    "loss": 0.9283761978149414,
+    "grad_norm": 0.3849687874317169,
+    "learning_rate": 3.332237841745898e-05,
+    "epoch": 1.953125,
+    "step": 125
+  },
+  {
+    "loss": 1.113053035736084,
+    "grad_norm": 0.4142734110355377,
+    "learning_rate": 2.9564590321322207e-05,
+    "epoch": 2.03125,
+    "step": 130
+  },
+  {
+    "loss": 0.9842248916625976,
+    "grad_norm": 0.44529953598976135,
+    "learning_rate": 2.5940702775459747e-05,
+    "epoch": 2.109375,
+    "step": 135
+  },
+  {
+    "loss": 0.9449721336364746,
+    "grad_norm": 0.3756776750087738,
+    "learning_rate": 2.2474460864709824e-05,
+    "epoch": 2.1875,
+    "step": 140
+  },
+  {
+    "loss": 1.0590093612670899,
+    "grad_norm": 0.4192875325679779,
+    "learning_rate": 1.9188576719953633e-05,
+    "epoch": 2.265625,
+    "step": 145
+  },
+  {
+    "loss": 0.9768091201782226,
+    "grad_norm": 0.5095818638801575,
+    "learning_rate": 1.6104580699624837e-05,
+    "epoch": 2.34375,
+    "step": 150
+  },
+  {
+    "loss": 1.038302516937256,
+    "grad_norm": 0.41709497570991516,
+    "learning_rate": 1.3242680314639993e-05,
+    "epoch": 2.421875,
+    "step": 155
+  },
+  {
+    "loss": 0.9975608825683594,
+    "grad_norm": 0.5563586354255676,
+    "learning_rate": 1.0621627821127289e-05,
+    "epoch": 2.5,
+    "step": 160
+  },
+  {
+    "loss": 0.9714397430419922,
+    "grad_norm": 0.8915637135505676,
+    "learning_rate": 8.25859734853645e-06,
+    "epoch": 2.578125,
+    "step": 165
+  },
+  {
+    "loss": 0.9948483467102051,
+    "grad_norm": 0.4391196370124817,
+    "learning_rate": 6.16907236823262e-06,
+    "epoch": 2.65625,
+    "step": 170
+  },
+  {
+    "loss": 0.9389057159423828,
+    "grad_norm": 0.4650712311267853,
+    "learning_rate": 4.366744239922998e-06,
+    "epoch": 2.734375,
+    "step": 175
+  },
+  {
+    "loss": 1.06390380859375,
+    "grad_norm": 0.4836062788963318,
+    "learning_rate": 2.8634225006782865e-06,
+    "epoch": 2.8125,
+    "step": 180
+  },
+  {
+    "loss": 1.008359718322754,
+    "grad_norm": 0.45215511322021484,
+    "learning_rate": 1.6689574843694433e-06,
+    "epoch": 2.890625,
+    "step": 185
+  },
+  {
+    "loss": 1.0110493659973145,
+    "grad_norm": 0.5408219695091248,
+    "learning_rate": 7.911757785462881e-07,
+    "epoch": 2.96875,
+    "step": 190
+  },
+  {
+    "loss": 0.911649227142334,
+    "grad_norm": 0.4599083364009857,
+    "learning_rate": 2.3582894166930268e-07,
+    "epoch": 3.046875,
+    "step": 195
+  },
+  {
+    "loss": 0.9673548698425293,
+    "grad_norm": 0.43304941058158875,
+    "learning_rate": 6.5558167183898955e-09,
+    "epoch": 3.125,
+    "step": 200
+  },
+  {
+    "train_runtime": 4256.6409,
+    "train_samples_per_second": 0.752,
+    "train_steps_per_second": 0.047,
+    "total_flos": 4.259313762009523e+16,
+    "train_loss": 1.142699921131134,
+    "epoch": 3.125,
+    "step": 200
+  }
+]