mirror of
				https://github.com/ggml-org/llama.cpp.git
				synced 2025-10-30 08:42:00 +00:00 
			
		
		
		
	scripts : fix pattern and get n_tokens in one go (#10221)
This commit is contained in:
		| @@ -23,8 +23,9 @@ CUR_PROMPT_CACHE="${CHAT_SAVE_DIR}/current-cache.bin" | ||||
| NEXT_PROMPT_FILE="${CHAT_SAVE_DIR}/next-prompt.txt" | ||||
| NEXT_PROMPT_CACHE="${CHAT_SAVE_DIR}/next-cache.bin" | ||||
|  | ||||
| SESSION_SIZE_MSG_PATTERN='main: session file matches [[:digit:]]+ / [[:digit:]]+' | ||||
| SAMPLE_TIME_MSG_PATTERN='sample time =[[:space:]]+[[:digit:]]+.[[:digit:]]+ ms /[[:space:]]+[[:digit:]]+' | ||||
| SESSION_AND_SAMPLE_PATTERN='main: session file matches [[:digit:]]+ / [[:digit:]]+'\ | ||||
| '|'\ | ||||
| 'sampling time =[[:space:]]+[[:digit:]]+.[[:digit:]]+ ms /[[:space:]]+[[:digit:]]+' | ||||
| SED_DELETE_MESSAGES="/^(${USER_NAME}:|${AI_NAME}:|\\.\\.\\.)/,\$d" | ||||
|  | ||||
| CTX_SIZE=2048 | ||||
| @@ -129,15 +130,12 @@ while read -e line; do | ||||
|  | ||||
|     printf ' ' | ||||
|  | ||||
|     # HACK get num tokens from debug message | ||||
|     # TODO get both messages in one go | ||||
|     if  ! session_size_msg="$(tail -n30 "$LOG" | grep -oE "$SESSION_SIZE_MSG_PATTERN")" || | ||||
|         ! sample_time_msg="$(tail -n10 "$LOG" | grep -oE "$SAMPLE_TIME_MSG_PATTERN")"; then | ||||
|     if ! session_and_sample_msg=$(tail -n30 "$LOG" | grep -oE "$SESSION_AND_SAMPLE_PATTERN"); then | ||||
|         echo >&2 "Couldn't get number of tokens from ./llama-cli output!" | ||||
|         exit 1 | ||||
|     fi | ||||
|  | ||||
|     n_tokens=$(($(cut -d/ -f2 <<<"$session_size_msg") + $(cut -d/ -f2 <<<"$sample_time_msg"))) | ||||
|     n_tokens=$(awk '{sum+=$1} END {print sum}' <<< "$(cut -d/ -f2 <<< "$session_and_sample_msg")") | ||||
|  | ||||
|     if ((n_tokens > CTX_ROTATE_POINT)); then | ||||
|         tail -c+$((n_prompt_len_pre + 1)) "$CUR_PROMPT_FILE" >>"$NEXT_PROMPT_FILE" | ||||
|   | ||||
		Reference in New Issue
	
	Block a user
	 haopeng
					haopeng