Updating pruned50-quant model

README.md

@@ -50,8 +50,8 @@ Model evaluation metrics and results.
 
 | Benchmark                                      | Metric        | Llama-2-7b-evolcodealpaca  | Llama-2-7b-pruned50-retrained-evolcodealpaca-quant-ds |
 |------------------------------------------------|---------------|-------------|-------------------------------|
-| [HumanEval](https://arxiv.org/abs/2107.03374)  | pass@1        | 32.03       | 32.2                          |
+| [HumanEval](https://arxiv.org/abs/2107.03374)  | pass@1        | 32.03       | 36.34                         |
 
 ## Help
 
-For further support, and discussions on these models and AI in general, join [Neural Magic's Slack Community](https://join.slack.com/t/discuss-neuralmagic/shared_invite/zt-q1a1cnvo-YBoICSIw3L1dmQpjBeDurQ)
+For further support, and discussions on these models and AI in general, join [Neural Magic's Slack Community](https://join.slack.com/t/discuss-neuralmagic/shared_invite/zt-q1a1cnvo-YBoICSIw3L1dmQpjBeDurQ)

config.json

@@ -1,5 +1,5 @@
 {
-  "_name_or_path": "/home/abhinav/src/llama-recipes/llama_7b_evol_codealpaca_sparse/sparse-sft_cerebras_50sp_llama_lr5e-4_epochs3_gradclip5.0_cosine-/combined/",
+  "_name_or_path": "neuralmagic/Llama-2-7b-pruned50-retrained-evolcodealpaca-quant-ds",
   "architectures": [
     "LlamaForCausalLM"
   ],
@@ -21,8 +21,9 @@
   "rope_scaling": null,
   "rope_theta": 10000.0,
   "tie_word_embeddings": false,
-  "torch_dtype": "float32",
-  "transformers_version": "1.7.0.20240312",
+  "tokenizer_class": "LlamaTokenizerFast",
+  "torch_dtype": "float16",
+  "transformers_version": "1.7.0.20240506",
   "use_cache": true,
   "vocab_size": 32000
 }

model-orig.onnx

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:16cb24f5b49e63173e59d9da7164a74fe968a51ddf20ef30a65dd73b87342390
-size 1220058
+oid sha256:d363691fa0eab6564db9d26c5603fd328d74a9860e2d73e962f9d000913181df
+size 1222751

model.data

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:48e44b79d03f06427edc1d451b31169c04966366fff80933784bb4d84df0f213
-size 7425272832
+oid sha256:aa7081720c95cd1692e42b5cc95b1fded0d9b7560058076bbc9168835a42372c
+size 7154772992

model.onnx

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:db93bc16d048df38acd6eaa934e5d1bb0beca2becc6023ea0ebcdfb090a697ee
-size 1205003
+oid sha256:7a6ad1f6b97d487e6966d5b7902b2aec84bf28771841d3c09d72bf73d75591e4
+size 1207696

tokenizer.json

@@ -2,7 +2,7 @@
   "version": "1.0",
   "truncation": {
     "direction": "Right",
-    "max_length": 2048,
+    "max_length": 384,
     "strategy": "LongestFirst",
     "stride": 0
   },

tokenizer_config.json

@@ -29,15 +29,11 @@
   "clean_up_tokenization_spaces": false,
   "eos_token": "</s>",
   "legacy": false,
-  "max_length": 2048,
   "model_max_length": 1000000000000000019884624838656,
   "pad_token": "</s>",
   "padding_side": "right",
   "sp_model_kwargs": {},
-  "stride": 0,
   "tokenizer_class": "LlamaTokenizer",
-  "truncation_side": "right",
-  "truncation_strategy": "longest_first",
   "unk_token": "<unk>",
   "use_default_system_prompt": false
 }