• Joined on 2024-05-31
tacit synced commits to refs/pull/4/merge at tacit/nanochat from mirror 2025-10-15 19:02:15 +00:00
190d9515d0 dont evaluate the sampling evals during SFT they are too slow. keep the multiple choice evals. delete unused imports
b8076dd367 fix bug in learning rate multiplier, it was ramping up instead of ramping down. see more in Issue #68. also add --dry_run option useful for experimentation
Compare 3 commits »
tacit synced commits to refs/pull/50/merge at tacit/nanochat from mirror 2025-10-15 19:02:15 +00:00
190d9515d0 dont evaluate the sampling evals during SFT they are too slow. keep the multiple choice evals. delete unused imports
b8076dd367 fix bug in learning rate multiplier, it was ramping up instead of ramping down. see more in Issue #68. also add --dry_run option useful for experimentation
Compare 3 commits »
tacit synced commits to refs/pull/49/merge at tacit/nanochat from mirror 2025-10-15 19:02:15 +00:00
190d9515d0 dont evaluate the sampling evals during SFT they are too slow. keep the multiple choice evals. delete unused imports
b8076dd367 fix bug in learning rate multiplier, it was ramping up instead of ramping down. see more in Issue #68. also add --dry_run option useful for experimentation
Compare 3 commits »
tacit synced commits to refs/pull/46/merge at tacit/nanochat from mirror 2025-10-15 19:02:15 +00:00
190d9515d0 dont evaluate the sampling evals during SFT they are too slow. keep the multiple choice evals. delete unused imports
b8076dd367 fix bug in learning rate multiplier, it was ramping up instead of ramping down. see more in Issue #68. also add --dry_run option useful for experimentation
Compare 3 commits »
tacit synced commits to refs/pull/59/head at tacit/nanochat from mirror 2025-10-15 19:02:15 +00:00
42b05eea7e Add guard against division by zero in chat_sft when num_tokens is 0
tacit synced commits to refs/pull/56/merge at tacit/nanochat from mirror 2025-10-15 19:02:15 +00:00
190d9515d0 dont evaluate the sampling evals during SFT they are too slow. keep the multiple choice evals. delete unused imports
b8076dd367 fix bug in learning rate multiplier, it was ramping up instead of ramping down. see more in Issue #68. also add --dry_run option useful for experimentation
Compare 3 commits »
tacit synced commits to refs/pull/54/merge at tacit/nanochat from mirror 2025-10-15 19:02:15 +00:00
190d9515d0 dont evaluate the sampling evals during SFT they are too slow. keep the multiple choice evals. delete unused imports
b8076dd367 fix bug in learning rate multiplier, it was ramping up instead of ramping down. see more in Issue #68. also add --dry_run option useful for experimentation
Compare 3 commits »
tacit synced commits to refs/pull/40/merge at tacit/nanochat from mirror 2025-10-15 19:02:15 +00:00
190d9515d0 dont evaluate the sampling evals during SFT they are too slow. keep the multiple choice evals. delete unused imports
b8076dd367 fix bug in learning rate multiplier, it was ramping up instead of ramping down. see more in Issue #68. also add --dry_run option useful for experimentation
Compare 3 commits »
tacit synced commits to refs/pull/43/merge at tacit/nanochat from mirror 2025-10-15 19:02:15 +00:00
190d9515d0 dont evaluate the sampling evals during SFT they are too slow. keep the multiple choice evals. delete unused imports
b8076dd367 fix bug in learning rate multiplier, it was ramping up instead of ramping down. see more in Issue #68. also add --dry_run option useful for experimentation
Compare 3 commits »
tacit synced commits to refs/pull/27/merge at tacit/nanochat from mirror 2025-10-15 19:02:14 +00:00
190d9515d0 dont evaluate the sampling evals during SFT they are too slow. keep the multiple choice evals. delete unused imports
b8076dd367 fix bug in learning rate multiplier, it was ramping up instead of ramping down. see more in Issue #68. also add --dry_run option useful for experimentation
Compare 3 commits »
tacit synced commits to refs/pull/38/merge at tacit/nanochat from mirror 2025-10-15 19:02:14 +00:00
190d9515d0 dont evaluate the sampling evals during SFT they are too slow. keep the multiple choice evals. delete unused imports
b8076dd367 fix bug in learning rate multiplier, it was ramping up instead of ramping down. see more in Issue #68. also add --dry_run option useful for experimentation
Compare 3 commits »
tacit synced commits to refs/pull/36/merge at tacit/nanochat from mirror 2025-10-15 19:02:14 +00:00
190d9515d0 dont evaluate the sampling evals during SFT they are too slow. keep the multiple choice evals. delete unused imports
b8076dd367 fix bug in learning rate multiplier, it was ramping up instead of ramping down. see more in Issue #68. also add --dry_run option useful for experimentation
Compare 3 commits »
tacit synced commits to refs/pull/35/merge at tacit/nanochat from mirror 2025-10-15 19:02:14 +00:00
190d9515d0 dont evaluate the sampling evals during SFT they are too slow. keep the multiple choice evals. delete unused imports
b8076dd367 fix bug in learning rate multiplier, it was ramping up instead of ramping down. see more in Issue #68. also add --dry_run option useful for experimentation
Compare 3 commits »
tacit synced commits to refs/pull/34/merge at tacit/nanochat from mirror 2025-10-15 19:02:14 +00:00
190d9515d0 dont evaluate the sampling evals during SFT they are too slow. keep the multiple choice evals. delete unused imports
b8076dd367 fix bug in learning rate multiplier, it was ramping up instead of ramping down. see more in Issue #68. also add --dry_run option useful for experimentation
Compare 3 commits »
tacit synced commits to refs/pull/32/merge at tacit/nanochat from mirror 2025-10-15 19:02:14 +00:00
190d9515d0 dont evaluate the sampling evals during SFT they are too slow. keep the multiple choice evals. delete unused imports
b8076dd367 fix bug in learning rate multiplier, it was ramping up instead of ramping down. see more in Issue #68. also add --dry_run option useful for experimentation
Compare 3 commits »
tacit synced commits to refs/pull/31/merge at tacit/nanochat from mirror 2025-10-15 19:02:14 +00:00
190d9515d0 dont evaluate the sampling evals during SFT they are too slow. keep the multiple choice evals. delete unused imports
b8076dd367 fix bug in learning rate multiplier, it was ramping up instead of ramping down. see more in Issue #68. also add --dry_run option useful for experimentation
Compare 3 commits »
tacit synced commits to refs/pull/30/merge at tacit/nanochat from mirror 2025-10-15 19:02:14 +00:00
190d9515d0 dont evaluate the sampling evals during SFT they are too slow. keep the multiple choice evals. delete unused imports
b8076dd367 fix bug in learning rate multiplier, it was ramping up instead of ramping down. see more in Issue #68. also add --dry_run option useful for experimentation
Compare 3 commits »
tacit synced commits to refs/pull/3/merge at tacit/nanochat from mirror 2025-10-15 19:02:14 +00:00
190d9515d0 dont evaluate the sampling evals during SFT they are too slow. keep the multiple choice evals. delete unused imports
b8076dd367 fix bug in learning rate multiplier, it was ramping up instead of ramping down. see more in Issue #68. also add --dry_run option useful for experimentation
Compare 3 commits »
tacit synced commits to refs/pull/24/merge at tacit/nanochat from mirror 2025-10-15 19:02:14 +00:00
190d9515d0 dont evaluate the sampling evals during SFT they are too slow. keep the multiple choice evals. delete unused imports
b8076dd367 fix bug in learning rate multiplier, it was ramping up instead of ramping down. see more in Issue #68. also add --dry_run option useful for experimentation
Compare 3 commits »
tacit synced commits to refs/pull/21/merge at tacit/nanochat from mirror 2025-10-15 19:02:14 +00:00
190d9515d0 dont evaluate the sampling evals during SFT they are too slow. keep the multiple choice evals. delete unused imports
b8076dd367 fix bug in learning rate multiplier, it was ramping up instead of ramping down. see more in Issue #68. also add --dry_run option useful for experimentation
Compare 3 commits »