As far as I can recall, both placental and marsupial mammals evolved in roughly the same time frame, approximately 130 to 136 million years ago in central asia, roughly china. As I understand their vector of transmission, they moved out from there into Europe and North America, from North America into South America, Antarctica and finally Australia. Essentially, spreading through the northern continent and then to the southern continent.
Turtledove's Atlantean territory would be right on this highway, so depending on the split, the Atlantean minicontinent could have had a viable collection of mammals at the time of the KT event.
But let's take another look at the geography here. What's interesting about Atlantis is that it is a tall minicontinent, but not a broad one. Taking a leaf from Jared Diamond, it strikes me that Atlantis would likely be a species diverse place.
Basically, stretching across a series of latitudes, it would tend to have a slightly different ratio of seasons and seasonal conditions all the way up and down. This would tend to encourage niche specialization.
Note that Atlantis is fairly narrow, so your miniclimates would be different on the east and west side as well, depending on ocean currents and wind and rain conditions.
Given a variety of distinct miniclimates, I would expect Atlantis to tend towards biological diversity, lots and lots of varieties of critter and plant. Which would probably make for pretty rapid evolution. Atlantis is a laboratory to promote rapid speciation.
On the other hand, Atlantis probably won't produce major megafauna. Not a big enough chunk of continuous habitat to produce megafauna. The best you'd get would be some land migrators, but I doubt it. While you'd have enough latitudes to specialize your climate, there wouldn't be enough meteorlogical space for them.
Let's see - Dinosaurs? No luck. I can't construct a plausible theory for the dinosaurs having Atlantis as a refuge.
Well, possibly. If Atlantean dinos evolved a wide diversity of small bodied forms, just maybe. But unlikely.
A mammal resurgence would be more likely. Not necessarily marsupials though, they seem to do better in more extreme and marginal environments. Believe it or not, Placental mammals colonized Australia at the same time as or earlier than Marsupials did. The Marsupials simply won the battle. They could handle the arid, biologically impoverished conditions better.
Atlantis, more likely a Placental homeland. But we could expect the local placentals to diversify into their own lineages. So no Canids, Felids, Ursines or Mustelids.
But probably something carnivorous. Marsupial Carnivores? Possibly. Big Bird Carnivores? Also possible. Also, resurgent crocodilian lines, including land crocs. Basically, the dog eat dog lifestyle seems to be open to all comers.
Herbivores? Likely complete domination by Placentals. I'm thinking predominantly forest as opposed to plains species. Browsers rather than grazers. Dear or Moose analogues. Possible Rodent analogues. Edentates?
With a bit of inspiration and enthusiasm, you could come up with some neat stuff.
Maybe even an intelligent species.