mirror of
https://github.com/ollama/ollama.git
synced 2025-03-20 06:42:41 +01:00
4879a234c4
* llama: wire up builtin runner This adds a new entrypoint into the ollama CLI to run the cgo built runner. On Mac arm64, this will have GPU support, but on all other platforms it will be the lowest common denominator CPU build. After we fully transition to the new Go runners more tech-debt can be removed and we can stop building the "default" runner via make and rely on the builtin always. * build: Make target improvements Add a few new targets and help for building locally. This also adjusts the runner lookup to favor local builds, then runners relative to the executable, and finally payloads. * Support customized CPU flags for runners This implements a simplified custom CPU flags pattern for the runners. When built without overrides, the runner name contains the vector flag we check for (AVX) to ensure we don't try to run on unsupported systems and crash. If the user builds a customized set, we omit the naming scheme and don't check for compatibility. This avoids checking requirements at runtime, so that logic has been removed as well. This can be used to build GPU runners with no vector flags, or CPU/GPU runners with additional flags (e.g. AVX512) enabled. * Use relative paths If the user checks out the repo in a path that contains spaces, make gets really confused so use relative paths for everything in-repo to avoid breakage. * Remove payloads from main binary * install: clean up prior libraries This removes support for v0.3.6 and older versions (before the tar bundle) and ensures we clean up prior libraries before extracting the bundle(s). Without this change, runners and dependent libraries could leak when we update and lead to subtle runtime errors.
97 lines
1.9 KiB
Go
97 lines
1.9 KiB
Go
package runner
|
|
|
|
import (
|
|
"strings"
|
|
)
|
|
|
|
func findStop(sequence string, stops []string) (bool, string) {
|
|
for _, stop := range stops {
|
|
if strings.Contains(sequence, stop) {
|
|
return true, stop
|
|
}
|
|
}
|
|
|
|
return false, ""
|
|
}
|
|
|
|
func containsStopSuffix(sequence string, stops []string) bool {
|
|
for _, stop := range stops {
|
|
for i := 1; i <= len(stop); i++ {
|
|
if strings.HasSuffix(sequence, stop[:i]) {
|
|
return true
|
|
}
|
|
}
|
|
}
|
|
|
|
return false
|
|
}
|
|
|
|
// truncateStop removes the provided stop string from pieces,
|
|
// returning the partial pieces with stop removed, including truncating
|
|
// the last piece if required (and signalling if this was the case)
|
|
func truncateStop(pieces []string, stop string) ([]string, bool) {
|
|
joined := strings.Join(pieces, "")
|
|
|
|
index := strings.Index(joined, stop)
|
|
if index == -1 {
|
|
return pieces, false
|
|
}
|
|
|
|
joined = joined[:index]
|
|
|
|
// Split truncated string back into pieces of original lengths
|
|
lengths := make([]int, len(pieces))
|
|
for i, piece := range pieces {
|
|
lengths[i] = len(piece)
|
|
}
|
|
|
|
var result []string
|
|
tokenTruncated := false
|
|
start := 0
|
|
for _, length := range lengths {
|
|
if start >= len(joined) {
|
|
break
|
|
}
|
|
|
|
end := start + length
|
|
if end > len(joined) {
|
|
end = len(joined)
|
|
tokenTruncated = true
|
|
}
|
|
result = append(result, joined[start:end])
|
|
start = end
|
|
}
|
|
|
|
return result, tokenTruncated
|
|
}
|
|
|
|
func incompleteUnicode(token string) bool {
|
|
incomplete := false
|
|
|
|
// check if there is incomplete UTF-8 character at the end
|
|
for i := 1; i < 5 && i <= len(token); i++ {
|
|
c := token[len(token)-i]
|
|
|
|
if (c & 0xc0) == 0x80 {
|
|
// continuation byte: 10xxxxxx
|
|
continue
|
|
}
|
|
|
|
if (c & 0xe0) == 0xc0 {
|
|
// 2-byte character: 110xxxxx ...
|
|
incomplete = i < 2
|
|
} else if (c & 0xf0) == 0xe0 {
|
|
// 3-byte character: 1110xxxx ...
|
|
incomplete = i < 3
|
|
} else if (c & 0xf8) == 0xf0 {
|
|
// 4-byte character: 11110xxx ...
|
|
incomplete = i < 4
|
|
}
|
|
|
|
// else 1-byte character or invalid byte
|
|
break
|
|
}
|
|
|
|
return incomplete
|
|
}
|