From: Georgi Gerganov <redacted>
Date: Mon, 17 Oct 2022 21:53:06 +0000 (+0300)
Subject: Fix EOT token handling
X-Git-Tag: upstream/1.7.4~1948
X-Git-Url: https://git.djapps.eu/?a=commitdiff_plain;h=cf67bfffa0acec5d09d9756d47e4333c57c92e8c;p=pkg%2Fggml%2Fsources%2Fwhisper.cpp

Fix EOT token handling

If it is the end of the audio, pick all sampled tokens.
Otherwise, print error message.
---

diff --git a/whisper.cpp b/whisper.cpp
index 98852781..ab229887 100644
--- a/whisper.cpp
+++ b/whisper.cpp
@@ -2445,9 +2445,12 @@ int whisper_full(
                 // end of text token
                 if (id == whisper_token_eot(ctx)) {
                     if (result_len == 0) {
-                        // TODO: figure out how to resolve this
-                        fprintf(stderr, "\n%s: failed to generate timestamp token - this should not happen\n\n", __func__);
-                        //result_len = i + 1;
+                        if (seek + seek_delta + 100 >= whisper_n_len(ctx)) {
+                            result_len = i + 1;
+                        } else {
+                            // TODO: figure out how to resolve this
+                            fprintf(stderr, "\n%s: failed to generate timestamp token - this should not happen\n\n", __func__);
+                        }
                     }
                     break;
                 }