Load parallel.cpp -f file.txt external prompt file#3416
Load parallel.cpp -f file.txt external prompt file#3416ggerganov merged 36 commits intoggml-org:masterfrom
Conversation
|
Thank you for the review. I have made all the changes.
Unfortunately I made a mess of updating this branch out of inexperience
with github processes so please ignore what I have pushed and I will do it
again today.
Is your preference for a new PR after this kind of updating?
…On Mon, Oct 2, 2023 at 12:17 PM Georgi Gerganov ***@***.***> wrote:
***@***.**** requested changes on this pull request.
------------------------------
On cmake_all.sh
<#3416 (comment)>:
Not needed
------------------------------
On ParallelQuestions.txt
<#3416 (comment)>:
Move to prompts change name to parallel-questions.txt
------------------------------
In common/common.h
<#3416 (comment)>:
> @@ -79,6 +79,7 @@ struct gpt_params {
std::string model_draft = ""; // draft model for speculative decoding
std::string model_alias = "unknown"; // model alias
std::string prompt = "";
+ std::string prompt_file = ""; // store for external prompt file name
⬇️ Suggested change
- std::string prompt_file = ""; // store for external prompt file name
+ std::string prompt_file = ""; // store the external prompt file name
------------------------------
In examples/parallel/parallel.cpp
<#3416 (comment)>:
> @@ -70,6 +72,22 @@ struct client {
std::vector<llama_token> tokens_prev;
};
+static void printDateTime() {
+ std::time_t currentTime = std::time(nullptr);
+ std::cout << "\n\033[35mRUN PARAMETERS as at \033[0m" << std::ctime(¤tTime);
We don't use std::cout in this project
------------------------------
In examples/parallel/parallel.cpp
<#3416 (comment)>:
> @@ -70,6 +72,22 @@ struct client {
std::vector<llama_token> tokens_prev;
};
+static void printDateTime() {
+ std::time_t currentTime = std::time(nullptr);
+ std::cout << "\n\033[35mRUN PARAMETERS as at \033[0m" << std::ctime(¤tTime);
+}
+
+// Define a split string function to ...
+static std::vector<std::string> splitString(const std::string& input, char delimiter) {
snake_case
—
Reply to this email directly, view it on GitHub
<#3416 (review)>,
or unsubscribe
<https://github.com/notifications/unsubscribe-auth/AGG22YKKD3B7MW56F3APRSTX5KPFRANCNFSM6AAAAAA5NU6KQE>
.
You are receiving this because you authored the thread.Message ID:
***@***.***>
|
No, please keep the existing PR. You may |
Thank you. All changes to <origin/load-parallel-prompt-file> now pushed to <origin/Update-load-parallel-prompt-file> which I hope I have done correctly this time. |
|
It's interesting to use the 100 questions in Since memory is critical it's worth noting the resources used and how parsimonious the system allocation is with `device.recommendedMaxWorkingSetSize': |
Could you please push your changes to the load-parallel-prompt-file branch so they appear here? |
…Update-load-parallel-prompt-file with requested changes
…/pudepiedj/llama.cpp into Update-load-parallel-prompt-file
OK this is what I did. I hope it's right. I've looked at the files in load-parallel-prompt-file and they appear to have been changed correctly. Please let me know if I have done something wrong (again)! |
…edj/llama.cpp into load-parallel-prompt-file
…edj/llama.cpp into load-parallel-prompt-file
…edj/llama.cpp into load-parallel-prompt-file
…example * 'master' of github.com:ggerganov/llama.cpp: kv cache slot search improvements (ggml-org#3493) prompts : fix editorconfig checks after ggml-org#3416 parallel : add option to load external prompt file (ggml-org#3416) server : reuse llama_sample_token common util (ggml-org#3494) llama : correct hparams comparison (ggml-org#3446) ci : fix xcodebuild destinations (ggml-org#3491) convert : update Falcon script for new HF config (ggml-org#3448) build : use std::make_tuple() for compatibility with older GCC versions (ggml-org#3488) common : process escape sequences in reverse prompts (ggml-org#3461) CLBlast: Fix handling of on-device tensor data server : fix incorrect num_tokens_predicted (ggml-org#3480) swift : disable ACCELERATE_NEW_LAPACK (ggml-org#3481) ci : add swift build via xcodebuild (ggml-org#3482)
|
Thank you. I appreciate your patience!
…On Fri, Oct 6, 2023 at 2:16 PM Georgi Gerganov ***@***.***> wrote:
Merged #3416 <#3416> into
master.
—
Reply to this email directly, view it on GitHub
<#3416 (comment)>, or
unsubscribe
<https://github.com/notifications/unsubscribe-auth/AGG22YMNL45H37DQSGKLLV3X6AAEFAVCNFSM6AAAAAA5NU6KQGVHI2DSMVQWIX3LMV45UABCJFZXG5LFIV3GK3TUJZXXI2LGNFRWC5DJN5XDWMJQGU3TKMJTGAZDGMQ>
.
You are receiving this because you authored the thread.Message ID:
***@***.***>
|
* Enable external file and add datestamp * Add name of external file at end * Upload ToK2024 * Delete ToK2024.txt * Experiments with jeopardy * Move ParallelQuestions to /proimpts and rename * Interim commit * Interim commit * Final revision * Remove trailing whitespace * remove cmake_all.sh * Remove cmake_all.sh * Changed .gitignore * Improved reporting and new question files. * Corrected typo * More LLM questions * Update LLM-questions.txt * Yet more LLM-questions * Remove jeopardy results file * Reinstate original jeopardy.sh * Update examples/parallel/parallel.cpp --------- Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>
This branch includes amendments to three files in
./llama.cpp/examplesnecessary to implement the external prompt file option-f file.txtthat arises in./bin/parallel --help. The three affected files are:Command-line code (second run example with
-ns 128):Example output from two different runs on M2 MAX 32GB and MacOS Sonoma 14.0 (omitting the initialisation):
Many lines omitted. This from the end of a run with
-ns 128.