Merge pull request #2295 from ArmDeveloperEcosystem/main

Production update

Author: pareenaverma

.wordlist.txt

@@ -4666,5 +4666,60 @@ crosh
 Sommelier
 chromeos
 linuxcontainers
-
+XPS
+NIC's 
+offlines
+passthrough
+SLOs
+Ker
+Rui
+SmartNICs
+selectedalt
+UIalt
+lpprojectubuntuarm
+RDV
+chiplet
+BMC
+upstreams
+rdv
+Initrd
+handoff
+ACPI
+PCRs
+MHU
+Handoff
+userland
+CXL
+DDR
+PHYs
+UCIe
+handoffs
+CCG
+CML
+Codespaces
+Cheng
+GDM
+LPI
+nsec
+shortcode
+BSON
+joedog
+Seige
+Antonov
+jwt
+kbs
+Nfpl
+ZjnAMjLk
+hCpeYsarnnGv
+kbs
+rvps
+xcbTMTBX
+CDH
+RVPS
+Attester
+attester
+ATtestation
+CoCo
+procedureS
+NIC’s
 

content/learning-paths/cross-platform/_example-learning-path/appendix-1-formatting.md

@@ -83,12 +83,26 @@ Specify that line_numbers are true in the following way:
 \`\`\`bash { line_numbers = "true" } \
 echo 'hello world' \
 echo ‘I am line two’ \
-\`\`\` 
+\`\`\`
+
+```bash { line_numbers = "true" }
+echo ‘hello world’
+echo ‘I am line two’
+```
+
+In some cases, the line numbering should not start from one but from another
+value, e.g. if the code excerpt is extracted from a larger file. Use the
+`line_start` attribute to achieve this:
 
-```bash { line_numbers = "true" } 
-echo ‘hello world’ 
-echo ‘I am line two’ 
-``` 
+\`\`\`bash { line_numbers = "true" line_start = "10" } \
+echo 'hello world' \
+echo ‘I am line two’ \
+\`\`\`
+
+```bash { line_numbers = "true" line_start = "10" }
+echo ‘hello world’
+echo ‘I am line eleven’
+```
 
 ### Output Lines
 
@@ -100,7 +114,7 @@ There are three ways you can specify command outputs in code:
 {{% notice Note %}}
 In each of the three situations, code marked as 'output' will:
 - not be copied when clicking the 'copy' button
-- not be highlightable by a cursor
+- not be highlighted by a cursor
 - appear slightly darker
 {{% /notice %}}
 

content/learning-paths/cross-platform/floating-point-rounding-errors/_index.md

@@ -1,6 +1,10 @@
 ---
 title: Explore floating-point differences between x86 and Arm
 
+draft: true
+cascade:
+    draft: true
+
 minutes_to_complete: 30
 
 who_is_this_for: This is an introductory topic for developers who are porting applications from x86 to Arm and want to understand how floating-point behavior differs between these architectures - particularly in the context of numerical consistency, performance, and debugging subtle bugs.

content/learning-paths/embedded-and-microcontrollers/training-inference-pytorch/_index.md

@@ -7,41 +7,41 @@ cascade:
 
 minutes_to_complete: 90
 
-who_is_this_for: This topic is for machine learning engineers, embedded AI developers, and researchers interested in deploying TinyML models for NLP on Arm-based edge devices using PyTorch and ExecuTorch. 
+who_is_this_for: This topic is for machine learning engineers, embedded AI developers, and researchers interested in deploying TinyML models for NLP on Arm-based edge devices using PyTorch and ExecuTorch.
 
-learning_objectives: 
+learning_objectives:
     - Train a custom CNN-based sentiment classification model implemented in PyTorch.
     - Optimize and convert the model using ExecuTorch for Arm-based edge devices.
     - Deploy and run inference on the Corstone-320 FVP.
 
 prerequisites:
-   - Basic knowledge of machine learning concepts. 
-   - It is advised to complete The Learning Path, [Introduction to TinyML on Arm using PyTorch and ExecuTorch](/learning-paths/embedded-and-microcontrollers/introduction-to-tinyml-on-arm) before starting this learning path. 
+   - Basic knowledge of machine learning concepts.
+   - It is advised to complete The Learning Path, [Introduction to TinyML on Arm using PyTorch and ExecuTorch](/learning-paths/embedded-and-microcontrollers/introduction-to-tinyml-on-arm) before starting this learning path.
    - Familiarity with Python and PyTorch.
    - A Linux host machine or VM running Ubuntu 22.04 or higher.
-   - An Arm license to run the examples on the Corstone-320 Fixed Virtual Platform (FVP), for hands-on deployment.  
+   - An Arm license to run the examples on the Corstone-320 Fixed Virtual Platform (FVP), for hands-on deployment.
 
 
 author: Dominica Abena O. Amanfo
 
 ### Tags
-skilllevels: Intermediate 
+skilllevels: Introductory
 subjects: ML
 armips:
     - Cortex-A
 tools_software_languages:
-    - tinyML 
-    - CNN 
+    - tinyML
+    - CNN
     - PyTorch
     - ExecuTorch
-    
+
 operatingsystems:
     - Linux
 
 
 further_reading:
     - resource:
-        title: Run Llama 3 on a Raspberry Pi 5 using ExecuTorch 
+        title: Run Llama 3 on a Raspberry Pi 5 using ExecuTorch
         link: /learning-paths/embedded-and-microcontrollers/rpi-llama3
         type: website
     - resource:

content/learning-paths/embedded-and-microcontrollers/uvprojx-conversion/_index.md

@@ -5,7 +5,7 @@ minutes_to_complete: 10
 
 who_is_this_for: This is a topic for users of µVision who want to migrate to the new project format (csolution) required by CMSIS-Toolbox.
 
-learning_objectives: 
+learning_objectives:
     - Import, convert, and build uvprojx-based projects in Keil Studio.
     - Convert uvprojx-based projects in µVision.
     - Convert and build uvprojx-based projects on the command line.
@@ -19,7 +19,7 @@ prerequisites:
 author: Christopher Seidl
 
 ### Tags
-skilllevels: Intermediate
+skilllevels: Advanced
 subjects: Performance and Architecture
 armips:
     - Cortex-M
@@ -43,7 +43,7 @@ further_reading:
         link: https://community.arm.com/arm-community-blogs/b/internet-of-things-blog/posts/keil-mdk-version-6
         type: blog
     - resource:
-        title: keil.arm.com 
+        title: keil.arm.com
         link: https://keil.arm.com
         type: website
 

content/learning-paths/laptops-and-desktops/self_hosted_cicd_github/_index.md

@@ -18,7 +18,7 @@ prerequisites:
 author: Dawid Borycki
 
 ### Tags
-skilllevels: Intermediate
+skilllevels: Introductory
 subjects: Migration to Arm
 armips:
     - Cortex-A

content/learning-paths/mobile-graphics-and-gaming/afrc/_index.md

@@ -5,7 +5,7 @@ minutes_to_complete: 25
 
 who_is_this_for: Software developers of Android applications and mobile games who are interested in learning how to enable Arm Fixed Rate Compression (AFRC) to improve performance.
 
-learning_objectives: 
+learning_objectives:
     - Query for fixed-rate compression support.
     - Specify what compression to use.
     - Verify that compression is applied.
@@ -18,7 +18,7 @@ prerequisites:
 author: Jose-Emilio Munoz-Lopez
 
 ### Tags
-skilllevels: Intermediate
+skilllevels: Advanced
 subjects: Graphics
 armips:
     - Mali

content/learning-paths/mobile-graphics-and-gaming/build-llama3-chat-android-app-using-executorch-and-xnnpack/5-run-benchmark-on-android.md

@@ -45,6 +45,7 @@ cmake -DCMAKE_TOOLCHAIN_FILE=$ANDROID_NDK/build/cmake/android.toolchain.cmake \
     -DEXECUTORCH_BUILD_KERNELS_CUSTOM=ON \
     -DEXECUTORCH_BUILD_KERNELS_LLM=ON \
     -DEXECUTORCH_BUILD_EXTENSION_LLM_RUNNER=ON \
+    -DEXECUTORCH_BUILD_EXTENSION_LLM=ON \
     -DEXECUTORCH_BUILD_EXTENSION_RUNNER_UTIL=ON \
     -DEXECUTORCH_XNNPACK_ENABLE_KLEIDI=ON \
     -DXNNPACK_ENABLE_ARM_BF16=OFF \
@@ -82,6 +83,10 @@ cmake --build cmake-out-android/examples/models/llama -j16 --config Release
 
 You should now have `llama_main` available for Android.
 
+{{% notice Note %}}
+If you notice that Gradle cannot find the Android SDK, add the sdk.dir path to executorch/extension/android/local.properties.
+{{% /notice %}}
+
 ## Run on Android via adb shell
 You will need an Arm-powered smartphone with the i8mm feature running Android, with 16GB of RAM. The following steps were tested on a Google Pixel 8 Pro phone.
 
@@ -103,7 +108,7 @@ You should see your device listed to confirm it is connected.
 
 ``` bash
 adb shell mkdir -p /data/local/tmp/llama
-adb push llama3_1B_kv_sdpa_xnn_qe_4_128_1024_embedding_4bit.pte /data/local/tmp/llama/
+adb push llama3_1B_kv_sdpa_xnn_qe_4_64_1024_embedding_4bit.pte /data/local/tmp/llama/
 adb push $HOME/.llama/checkpoints/Llama3.2-1B-Instruct/tokenizer.model /data/local/tmp/llama/
 adb push cmake-out-android/examples/models/llama/llama_main /data/local/tmp/llama/
 ```
@@ -114,49 +119,53 @@ adb push cmake-out-android/examples/models/llama/llama_main /data/local/tmp/llam
 Use the Llama runner to execute the model on the phone with the `adb` command:
 
 ``` bash
-adb shell "cd /data/local/tmp/llama && ./llama_main --model_path llama3_1B_kv_sdpa_xnn_qe_4_64_1024_embedding_4bit.pte --tokenizer_path tokenizer.model --prompt "<|start_header_id|>system<|end_header_id|>\nYour name is Cookie. you are helpful, polite, precise, concise, honest, good at writing. You always give precise and brief answers up to 32 words<|eot_id|><|start_header_id|>user<|end_header_id|>\nHey Cookie! how are you today?<|eot_id|><|start_header_id|>assistant<|end_header_id|>" --warmup=1 --cpu_threads=5
+adb shell "cd /data/local/tmp/llama && ./llama_main --model_path llama3_1B_kv_sdpa_xnn_qe_4_64_1024_embedding_4bit.pte --tokenizer_path tokenizer.model --prompt "<|start_header_id|>system<|end_header_id|>\nYour name is Cookie. you are helpful, polite, precise, concise, honest, good at writing. You always give precise and brief answers up to 32 words<|eot_id|><|start_header_id|>user<|end_header_id|>\nHey Cookie! how are you today?<|eot_id|><|start_header_id|>assistant<|end_header_id|>" --warmup=1 --cpu_threads=5"
 ```
 
 The output should look something like this.
 
 ```
-I 00:00:00.003316 executorch:main.cpp:69] Resetting threadpool with num threads = 5
-I 00:00:00.009329 executorch:runner.cpp:59] Creating LLaMa runner: model_path=llama3_1B_kv_sdpa_xnn_qe_4_64_1024_embedding_4bit.pte, tokenizer_path=tokenizer.model
-I 00:00:03.569399 executorch:runner.cpp:88] Reading metadata from model
-I 00:00:03.569451 executorch:runner.cpp:113] Metadata: use_sdpa_with_kv_cache = 1
-I 00:00:03.569455 executorch:runner.cpp:113] Metadata: use_kv_cache = 1
-I 00:00:03.569459 executorch:runner.cpp:113] Metadata: get_vocab_size = 128256
-I 00:00:03.569461 executorch:runner.cpp:113] Metadata: get_bos_id = 128000
-I 00:00:03.569464 executorch:runner.cpp:113] Metadata: get_max_seq_len = 1024
-I 00:00:03.569466 executorch:runner.cpp:113] Metadata: enable_dynamic_shape = 1
-I 00:00:03.569469 executorch:runner.cpp:120] eos_id = 128009
-I 00:00:03.569470 executorch:runner.cpp:120] eos_id = 128001
-I 00:00:03.569471 executorch:runner.cpp:120] eos_id = 128006
-I 00:00:03.569473 executorch:runner.cpp:120] eos_id = 128007
-I 00:00:03.569475 executorch:runner.cpp:168] Doing a warmup run...
-I 00:00:03.838634 executorch:text_prefiller.cpp:53] Prefill token result numel(): 128256
- 
-I 00:00:03.892268 executorch:text_token_generator.h:118]
+I tokenizers:regex.cpp:27] Registering override fallback regex
+I 00:00:00.003288 executorch:main.cpp:87] Resetting threadpool with num threads = 5
+I 00:00:00.006393 executorch:runner.cpp:44] Creating LLaMa runner: model_path=llama3_1B_kv_sdpa_xnn_qe_4_64_1024_embedding_4bit.pte, tokenizer_path=tokenizer.model
+E tokenizers:hf_tokenizer.cpp:60] Error parsing json file: [json.exception.parse_error.101] parse error at line 1, column 1: syntax error while parsing value - invalid literal; last read: 'I'
+I 00:00:00.131486 executorch:llm_runner_helper.cpp:57] Loaded TikToken tokenizer
+I 00:00:00.131525 executorch:llm_runner_helper.cpp:167] Reading metadata from model
+I 00:00:00.186538 executorch:llm_runner_helper.cpp:110] Metadata: use_sdpa_with_kv_cache = 1
+I 00:00:00.186574 executorch:llm_runner_helper.cpp:110] Metadata: use_kv_cache = 1
+I 00:00:00.186578 executorch:llm_runner_helper.cpp:110] Metadata: get_max_context_len = 1024
+I 00:00:00.186584 executorch:llm_runner_helper.cpp:110] Metadata: get_max_seq_len = 1024
+I 00:00:00.186588 executorch:llm_runner_helper.cpp:110] Metadata: enable_dynamic_shape = 1
+I 00:00:00.186596 executorch:llm_runner_helper.cpp:140] eos_id = 128009
+I 00:00:00.186597 executorch:llm_runner_helper.cpp:140] eos_id = 128001
+I 00:00:00.186599 executorch:llm_runner_helper.cpp:140] eos_id = 128006
+I 00:00:00.186600 executorch:llm_runner_helper.cpp:140] eos_id = 128007
+I 00:00:01.086570 executorch:text_llm_runner.cpp:89] Doing a warmup run...
+I 00:00:01.087836 executorch:text_llm_runner.cpp:152] Max new tokens resolved: 128, given start_pos 0, num_prompt_tokens 54, max_context_len 1024
+I 00:00:01.292740 executorch:text_prefiller.cpp:93] Prefill token result numel(): 128256
+
+I 00:00:02.264371 executorch:text_token_generator.h:123]
 Reached to the end of generation
-I 00:00:03.892281 executorch:runner.cpp:267] Warmup run finished!
-I 00:00:03.892286 executorch:runner.cpp:174] RSS after loading model: 1269.445312 MiB (0 if unsupported)
-<|start_header_id|>system<|end_header_id|>\nYour name is Cookie. you are helpful, polite, precise, concise, honest, good at writing. You always give precise and brief answers up to 32 words<|eot_id|><|start_header_id|>user<|end_header_id|>\nHey Cookie! how are you today?<|eot_id|><|start_header_id|>assistant<|end_header_id|>I 00:00:04.076905 executorch:text_prefiller.cpp:53] Prefill token result numel(): 128256
- 
- 
-I 00:00:04.078027 executorch:runner.cpp:243] RSS after prompt prefill: 1269.445312 MiB (0 if unsupported)
-I'm doing great, thanks! I'm always happy to help, communicate, and provide helpful responses. I'm a bit of a cookie (heh) when it comes to delivering concise and precise answers. What can I help you with today?<|eot_id|>
-I 00:00:05.399304 executorch:text_token_generator.h:118]
+I 00:00:02.264379 executorch:text_llm_runner.cpp:209] Warmup run finished!
+I 00:00:02.264384 executorch:text_llm_runner.cpp:95] RSS after loading model: 1122.187500 MiB (0 if unsupported)
+I 00:00:02.264624 executorch:text_llm_runner.cpp:152] Max new tokens resolved: 74, given start_pos 0, num_prompt_tokens 54, max_context_len 1024
+<|start_header_id|>system<|end_header_id|>\nYour name is Cookie. you are helpful, polite, precise, concise, honest, good at writing. You always give precise and brief answers up to 32 words<|eot_id|><|start_header_id|>user<|end_header_id|>\nHey Cookie! how are you today?<|eot_id|><|start_header_id|>assistant<|end_header_id|>I 00:00:02.394162 executorch:text_prefiller.cpp:93] Prefill token result numel(): 128256
+
+
+I 00:00:02.394373 executorch:text_llm_runner.cpp:179] RSS after prompt prefill: 1122.187500 MiB (0 if unsupported)
+I'm doing great, thanks for asking! I'm always ready to help, whether it's answering a question or providing a solution. What can I help you with today?<|eot_id|>
+I 00:00:03.072966 executorch:text_token_generator.h:123]
 Reached to the end of generation
- 
-I 00:00:05.399314 executorch:runner.cpp:257] RSS after finishing text generation: 1269.445312 MiB (0 if unsupported)
-PyTorchObserver {"prompt_tokens":54,"generated_tokens":51,"model_load_start_ms":1710296339487,"model_load_end_ms":1710296343047,"inference_start_ms":1710296343370,"inference_end_ms":1710296344877,"prompt_eval_end_ms":1710296343556,"first_token_ms":1710296343556,"aggregate_sampling_time_ms":49,"SCALING_FACTOR_UNITS_PER_SECOND":1000}
-I 00:00:04.530945 executorch:stats.h:108] 	Prompt Tokens: 54    Generated Tokens: 69
-I 00:00:04.530947 executorch:stats.h:114] 	Model Load Time:		1.196000 (seconds)
-I 00:00:04.530949 executorch:stats.h:124] 	Total inference time:		1.934000 (seconds)		 Rate: 	35.677353 (tokens/second)
-I 00:00:04.530952 executorch:stats.h:132] 		Prompt evaluation:	0.176000 (seconds)		 Rate: 	306.818182 (tokens/second)
-I 00:00:04.530954 executorch:stats.h:143] 		Generated 69 tokens:	1.758000 (seconds)		 Rate: 	39.249147 (tokens/second)
-I 00:00:04.530956 executorch:stats.h:151] 	Time to first generated token:	0.176000 (seconds)
-I 00:00:04.530959 executorch:stats.h:158] 	Sampling time over 123 tokens:	0.067000 (seconds)
+
+I 00:00:03.072972 executorch:text_llm_runner.cpp:199] RSS after finishing text generation: 1122.187500 MiB (0 if unsupported)
+PyTorchObserver {"prompt_tokens":54,"generated_tokens":36,"model_load_start_ms":1756473387815,"model_load_end_ms":1756473388715,"inference_start_ms":1756473389893,"inference_end_ms":1756473390702,"prompt_eval_end_ms":1756473390023,"first_token_ms":1756473390023,"aggregate_sampling_time_ms":22,"SCALING_FACTOR_UNITS_PER_SECOND":1000}
+I 00:00:03.072993 executorch:stats.h:108] 	Prompt Tokens: 54    Generated Tokens: 36
+I 00:00:03.072995 executorch:stats.h:114] 	Model Load Time:		0.900000 (seconds)
+I 00:00:03.072996 executorch:stats.h:124] 	Total inference time:		0.809000 (seconds)  Rate: 	44.499382 (tokens/second)
+I 00:00:03.072998 executorch:stats.h:132] 	Prompt evaluation:	0.130000 (seconds)          Rate: 	415.384615 (tokens/second)
+I 00:00:03.073000 executorch:stats.h:143] 	Generated 36 tokens:	0.679000 (seconds)      Rate: 	53.019146 (tokens/second)
+I 00:00:03.073002 executorch:stats.h:151] 	Time to first generated token:	0.130000 (seconds)
+I 00:00:03.073004 executorch:stats.h:158] 	Sampling time over 90 tokens:	0.022000 (seconds)
 ```
 
 You have successfully run the Llama 3.1 1B Instruct model on your Android smartphone with ExecuTorch using KleidiAI kernels.